11992 Subject: Re: Determining Limits of an Arbitrarily Rotated Ellispe > Anyone have the derivation for the equations to determine the X and > Y-axis limits (mins and maxes) of an arbitrarily rotated ellipse in an > X-Y plane? Currently given is the axis-alligned equation of the > ellipse and the angle of rotation to the X-axis. Conversly, the > generalized ellipse equation can be derived, too. > Thanks in advance. > A paper which derives the equations for determining these values for a > generalized ellipse can be found at > http://www.crbond.com/papers/ellipse.pdf . > This paper was written to present a method for scan converting ellipses at > arbitrary angles. > There are two things you must never attempt to prove: the unprovable -- > and the obvious. > Perfecto! :o) Thanks so much, C Subject: More glossary items (was: Extending Python Syntax with @) Originator: claird@lairds.com (Cameron Laird) >months. I took five years of calculus in college, and I still don't >see the connection between lambda functions and calculus. They're both about getting results. Serious. What you learned as calculus was earlier called the analytic calculus or the differential and integral calculus. It's about getting results (very roughly speaking) that have to do with smooth properties (of shapes and motions, let's say). Lambda is entirely different--except that it's also a calculus, that is, largely expressible algorithmically. Lambda happens to be about computations. There are other calculi--quaternion calculus and vector calculus are two not-too-uncommonly-heard ones. Lambda and the one you learned in college courses are much the most often used in such references. -- Business: http://www.Phaseit.net Subject: Re: More glossary items > There are other calculi--quaternion calculus and vector > calculus are two not-too-uncommonly-heard ones. Indeed, in mathematics the word calculus really just means a method of calculating something. It's actually the Latin word for stone, presumably dating back to the time when people used stones as counters when doing arithmetic. -- http://www.cosc.canterbury.ac.nz/~greg Subject: Root to One This attempts to solve polynomials. Eqn of plane, a[1]x[1]+a[2]x[2]+a[3]x[3]+a[4]x[4]+a[5]x[5]+a[6]x[6]+a[7]x[7] +a[0]=0 N=(a[1],a[2],a[3],a[4],a[5],a[6],a[7]) normal T=(t,t^2,t^3,t^4,t^5,t^6,t^7) space curve N*T + a[0]=0 7th degree polynomial or a[1]t+a[2]t^2+a[3]t^3+a[4]t^4+a[5]t^5+a[6]t^6+a[7]t^7+a[0]=0 T=(-a[0]/|N|^2)N This is shown by N*T = (-a[0]/|N|^2)N*N decoding, t =(-a[0]a[1])/|N|^2 = x[1] t^2=(-a[0]a[2])/|N|^2 = x[2] t^3=(-a[0]a[3])/|N|^2 = x[3] t^4=(-a[0]a[4])/|N|^2 = x[4] t^5=(-a[0]a[5])/|N|^2 = x[5] t^6=(-a[0]a[6])/|N|^2 = x[6] t^7=(-a[0]a[7])/|N|^2 = x[7] substituting these in the equation below, t= 2{ 2{ 2{ 2{ 2{ 2{x[7]}^(6/7)-x[6] }^(5/6)-x[5] }^(4/5)-x[4] }^(3/4)-x[3] }^(2/3)-x[2] }^(1/2)-x[1] solve, x^3+2x^2-3x+1=0 a[1]=-3 a[2]= 2 a[3]= 1 a[0]= 1 |N|^2 = 9+4+1=15 t = -a[0]a[1]/|N|^2 = 3/15 = x[1] t^2= -a[0]a[2]/|N|^2 = -2/15 = x[2] t^3= -a[0]a[3]/|N|^2 = -1/15 = x[3] t = 2{2{x[3]}^(2/3)-x[2]}^(1/2)-x[1] x = 2{2{-1/15}^(2/3)-(-2/15)}^(1/2)-3/15 x = .540+/-i(.183) , -3.079 roots from calculator Angle x subtends known arclength A and chord B (B/A)(x/2)=sin(x/2) trigonometric representation sin(x/2)=x/2-(x/2)^3/3!+(x/2)^5/5!-... Maclaurin sine B/A=1-(x/2)^2/3!+(x/2)^4/5!-... cancel x/2 t=x/2 x=2t a[odd]=0 a[0] =1-B/A a[2] = -1/3! a[4] = 1/5! a[6] = -1/7! a[8] = 1/9! a[10]= -1/11! a[12]= 1/13! a[14]= -1/15! |N|^2 = (-1/3!)^2+(1/5!)^2+(-1/7!)^2+...+(-1/15!)^2 x[2] =-a[0]a[ 2]/|N|^2= -(1-B/A)(-1/3! )/|N|^2 x[4] =-a[0]a[ 4]/|N|^2= -(1-B/A)( 1/5! )/|N|^2 x[6] =-a[0]a[ 6]/|N|^2= -(1-B/A)(-1/7! )/|N|^2 . . x[14]=-a[0]a[14]/|N|^2= -(1-B/A)(-1/15!)/|N|^2 (x/2)^2 = 2{ 2{ 2{ 2{ 2{ 2{x[14]}^(12/14)-x[12] }^(10/12)-x[10] }^( 8/10)-x[ 8] }^( 6/8 )-x[ 6] }^( 4/6 )-x[ 4] }^( 2/4 )-x[ 2] For small angles, B=A and x=0 Subject: Roots to Polynomials Eqn of plane, a[1]x[1]+a[2]x[2]+a[3]x[3]+a[4]x[4]+a[5]x[5]+a[6]x[6]+a[7]x[7] +a[0]=0 N=(a[1],a[2],a[3],a[4],a[5],a[6],a[7]) normal T=(t,t^2,t^3,t^4,t^5,t^6,t^7) space curve N*T + a[0]=0 7th degree polynomial or a[1]t+a[2]t^2+a[3]t^3+a[4]t^4+a[5]t^5+a[6]t^6+a[7]t^7+a[0]=0 T=(-a[0]/|N|^2)N This is shown by N*T = (-a[0]/|N|^2)N*N decoding, t =(-a[0]a[1])/|N|^2 = x[1] t^2=(-a[0]a[2])/|N|^2 = x[2] t^3=(-a[0]a[3])/|N|^2 = x[3] t^4=(-a[0]a[4])/|N|^2 = x[4] t^5=(-a[0]a[5])/|N|^2 = x[5] t^6=(-a[0]a[6])/|N|^2 = x[6] t^7=(-a[0]a[7])/|N|^2 = x[7] substituting these in the equation below, t= 2{ 2{ 2{ 2{ 2{ 2{x[7]}^(6/7)-x[6] }^(5/6)-x[5] }^(4/5)-x[4] }^(3/4)-x[3] }^(2/3)-x[2] }^(1/2)-x[1] solve, x^3+2x^2-3x+1=0 a[1]=-3 a[2]= 2 a[3]= 1 a[0]= 1 |N|^2 = 9+4+1=15 t = -a[0]a[1]/|N|^2 = 3/15 = x[1] t^2= -a[0]a[2]/|N|^2 = -2/15 = x[2] t^3= -a[0]a[3]/|N|^2 = -1/15 = x[3] t = 2{2{x[3]}^(2/3)-x[2]}^(1/2)-x[1] x = 2{2{-1/15}^(2/3)-(-2/15)}^(1/2)-3/15 x = .540+/-i(.183) , -3.079 roots from calculator Angle x subtends known arclength A and chord B (B/A)(x/2)=sin(x/2) trigonometric representation sin(x/2)=x/2-(x/2)^3/3!+(x/2)^5/5!-... Maclaurin sine B/A=1-(x/2)^2/3!+(x/2)^4/5!-... cancel x/2 t=x/2 x=2t a[odd]=0 a[0] =1-B/A a[2] = -1/3! a[4] = 1/5! a[6] = -1/7! a[8] = 1/9! a[10]= -1/11! a[12]= 1/13! a[14]= -1/15! |N|^2 = (-1/3!)^2+(1/5!)^2+(-1/7!)^2+...+(-1/15!)^2 x[2] =-a[0]a[ 2]/|N|^2= -(1-B/A)(-1/3! )/|N|^2 x[4] =-a[0]a[ 4]/|N|^2= -(1-B/A)( 1/5! )/|N|^2 x[6] =-a[0]a[ 6]/|N|^2= -(1-B/A)(-1/7! )/|N|^2 . . x[14]=-a[0]a[14]/|N|^2= -(1-B/A)(-1/15!)/|N|^2 (x/2)^2 = 2{ 2{ 2{ 2{ 2{ 2{x[14]}^(12/14)-x[12] }^(10/12)-x[10] }^( 8/10)-x[ 8] }^( 6/8 )-x[ 6] }^( 4/6 )-x[ 4] }^( 2/4 )-x[ 2] Subject: Re: Sum, Over Some Prime Multiples, Is Always m > Let P be any subset of the primes. > (example: P contains all primes of even-index, > or those congruent to 1 (mod 6).) > Let A be the set of positive integers: including 1 and every positive > integer which is a multiple of only the primes in P, and A contains > no member which is divisible by a prime not in P. > Let B be the set of positive integers: including 1 and every positive > integer which is a multiple of only the primes *not* in P, and B > contains no member which is divisible by a prime *in* P. > (Yes, there are positive integers in neither set, as long as P does > not contain every prime.) > Let g(x) be the number of distinct elements of A which are <= x, > for x = a positive real. > So then, for m = any positive integer: > --- > / g(m/k) > --- > k=elements of B, k <= m > always = m. > In linear-mode: > sum{k=elements of B, k<=m} g(m/k) > always equals m. > (Right?) > A consequence (which probably is trivially shown by a more direct method) of the above: If, for a given P (and A), limit{m->oo} g(m)/m = x exists and is finite nonzero, then sum{k=elements of B} 1/k ( = product{p=primes not in P} 1/(1 -1/p) ) converges to 1/x. And so, from this we see that sum{p=primes not in P} 1/p converges if the limit of g(m)/m approaches a limit which is finite nonzero. But...I am being VERY nonrigorous above, and could very well be wrong. Have I posted truth anyway? Subject: Re: Sum, Over Some Prime Multiples, Is Always m > Let P be any subset of the primes. > (example: P contains all primes of even-index, > or those congruent to 1 (mod 6).) > Let A be the set of positive integers: including 1 and every positive > integer which is a multiple of only the primes in P, and A contains > no member which is divisible by a prime not in P. > Let B be the set of positive integers: including 1 and every positive > integer which is a multiple of only the primes *not* in P, and B > contains no member which is divisible by a prime *in* P. > (Yes, there are positive integers in neither set, as long as P does > not contain every prime.) > Let g(x) be the number of distinct elements of A which are <= x, > for x = a positive real. > So then, for m = any positive integer: > --- > / g(m/k) > --- > k=elements of B, k <= m > always = m. > In linear-mode: > sum{k=elements of B, k<=m} g(m/k) > always equals m. > (Right?) > > A consequence (which probably is trivially shown by a more direct > method) of the above: > If, for a given P (and A), > limit{m->oo} g(m)/m = x > exists and is finite nonzero, then > sum{k=elements of B} 1/k > ( = product{p=primes not in P} 1/(1 -1/p) ) > > converges to 1/x. > And so, from this we see that > sum{p=primes not in P} 1/p converges > if the limit of g(m)/m approaches a limit which is finite nonzero. > But...I am being VERY nonrigorous above, and could very well be wrong. > Have I posted truth anyway? > So, unrigorously again, If sum{p=primes in P} 1/p AND sum{p=primes not in P} 1/p both diverge, then: the percentage of all positive integers neither in A nor in B approaches 100%. In other words, if j(m) = number of positive integers <= m which are not in A nor in B (ie. number of positive integers <= m which are divisible by *both* primes in P and in not-P), then limit{m->oo} j(m)/m = 1 IF the 2 prime-reciprocal series above *both* diverge. Even though this result can be shown, shown if true, without the use of the result in this thread's original post, it is directly related to this result. Subject: Re: Determining the Equation of a Shadow Cast onto a Surface <462e18e1.0403111517.71dd612@posting.google.com >> Given a flat circular disk (high-gain antenna), a planar surface (a >> solar array), and an arbitrary vector to a point light source (the >> Sun), I would like to determine the equation of the resultant shadow >> (which would be an ellipse) that would be cast onto the plane. >Oops, I posted this in the wrong place. >As additional information, the solar array is a group of maybe 10,000 >individual solar cells, and this analysis is to determine the power >output reduction over the entire array if a portion of those cells are >shadowed from the Sun by the high-gain antenna. So, certain >assumptions can simplify this analysis. Fringes can be ignored (a few >cells won't degrade the power output significantly). The Sun's disk >size can be ignored. Is this for visualization or for engineering? The reduction in power is going to depend on the sine of the angle of the disk with respect to the sun's rays and percentage of the shadow which falls on the array, multiplied by the area of the disk. I don't think you necessarily need to know the shape (equation) of the shadow on the array except to calculate the percentage which falls on the array. (Of course, if the array is not perpendicular to the direction of the sun, the output will further be reduced by the sine of that angle...) >Given knowledge of the physical characteristics of this spacecraft >(geometries and sizes), locations and orientations of the articulable >solar array and high-gain antenna, and solar system ephemerous, I was >planning on using an orthoganal projection approach to model the >shadow cast on the array. Orthagonal projection of the disk onto the array will only happen when the array is perpendicular to the sun's rays. Is that always going to be the case? >Yeah, I realize the shadow might project off to a parabola with some >points of the disk at infinity if the Sun is low enough (I belive I >could use some sort of projective transformation). Assuming the sun is larger than the disk, I don't think you need to worry about that! The above may not (yet) represent the opinions of my employer. Subject: Re: Determining the Equation of a Shadow Cast onto a Surface >> Given a flat circular disk (high-gain antenna), a planar surface (a >> solar array), and an arbitrary vector to a point light source (the >> Sun), I would like to determine the equation of the resultant shadow >> (which would be an ellipse) that would be cast onto the plane. >Oops, I posted this in the wrong place. >As additional information, the solar array is a group of maybe 10,000 >individual solar cells, and this analysis is to determine the power >output reduction over the entire array if a portion of those cells are >shadowed from the Sun by the high-gain antenna. So, certain >assumptions can simplify this analysis. Fringes can be ignored (a few >cells won't degrade the power output significantly). The Sun's disk >size can be ignored. > Is this for visualization or for engineering? Enginnering > The reduction in power is going to depend on the sine of the angle of the > disk with respect to the sun's rays and percentage of the shadow which falls > on the array, multiplied by the area of the disk. I don't think you > necessarily need to know the shape (equation) of the shadow on the array > except to calculate the percentage which falls on the array. (Of course, if > the array is not perpendicular to the direction of the sun, the output will > further be reduced by the sine of that angle...) It's a little more complicated than this. The individual solar cells are grouped in what are called strings of 16 cells. The input to the power prediction software is as follows: 1) The number of strings with 0 cells in shadow 2) The number of strings with 1 cell in shadow . . . 17) The number of strings with all 16 cells in shadow 18) The incedence angle between the solar array normal and the Sun vector Thus, my algorithm first isolates a bounding square on the array in which the shadow is cast (the minimum and maximum limits of the shadow ellipse in solar array coordinates), and then determines which cells in that square are in the elliptical shadow. The arrays will be normal to the Sun during most of the ops, however, a constraint is that we do not loose contact, so the HGA pointing at the Earth is given higher priority. This could result in some off-pointing of the arrays. >Given knowledge of the physical characteristics of this spacecraft >(geometries and sizes), locations and orientations of the articulable >solar array and high-gain antenna, and solar system ephemerous, I was >planning on using an orthoganal projection approach to model the >shadow cast on the array. > Orthagonal projection of the disk onto the array will only happen when the > array is perpendicular to the sun's rays. Is that always going to be the > case? >Yeah, I realize the shadow might project off to a parabola with some >points of the disk at infinity if the Sun is low enough (I belive I >could use some sort of projective transformation). > Assuming the sun is larger than the disk, I don't think you need to worry > about that! > > The above may not (yet) represent the opinions of my employer. Subject: Re: Question about Ox Metrics & Monte Carlo Simulations >I have a programming assignment I need to work on that involves the Ox >Metrics programming environment. Is anyone in this group familiar with >programming in Ox for the purpose of Monte Carlo simulations (for an >econometrics assignment)? If so, what expierience have you had with >this or do you have any recommendations for doing Monte Carlo >simulations in general? Monte Carlo simulations model many cases using random input variables in order to generate expected statistics. For this to work, it's important that: * Your random input variables have a distribution which matches reality * Your model is accurate * You run a large number of cases (tip: Do 2 or more runs with different random seeds. If the results are different, increase the number of cases per run. As the number of cases increases, the results should converge to the same set of values for any random seed.) Sorry I know nothing about Ox. The above may not (yet) represent the opinions of my employer. Subject: Re: Want ro receive A+ paper? whatever happened to bullying people into doing your homework? Subject: Could someone explain this... I am trying to break my way into Calculus, but I need some help in understanding limits, (which I know I need to know to move on...) Could someone please explain how to calculate a limit at Infinity? Subject: Re: Could someone explain this... > I am trying to break my way into Calculus, but I need some help in > understanding limits, (which I know I need to know to move on...) > Could someone please explain how to calculate a limit at Infinity? Intuitively, the limit as a function f goes to infinity is a number L that f gets arbitrarily close to as the domain of the function gets arbitrarily large. For example, f(x) = x. As x gets very large, f(x) gets very large, so we say the limit is infinity. Another example: f(x) = 1/x. As x gets very large, f(x) gets very close to 0, so the limit is 0. Perhaps if you post some of the specific problems you're having difficulty with, we can offer more specific help. Are you working from a book? Subject: Re: Could someone explain this... >> I am trying to break my way into Calculus, but I need some help in >> understanding limits, (which I know I need to know to move on...) > >> Could someone please explain how to calculate a limit at Infinity? >Intuitively, the limit as a function f goes to infinity is a number L >that f gets arbitrarily close to as the domain of the function gets >arbitrarily large. >For example, f(x) = x. As x gets very large, f(x) gets very large, so we >say the limit is infinity. >Another example: f(x) = 1/x. As x gets very large, f(x) gets very close >to 0, so the limit is 0. >Perhaps if you post some of the specific problems you're having >difficulty with, we can offer more specific help. Are you working from a >book? In calculus, I'd also expect l'Hopital's rule to be covered. You have to understand derivatives before you can use it. If that value of a function is indeterminate, that is 0/0 or infinity/infinity, you can take the derivative (differential) of the numerator and denominator. For example, if f(x) = 2x / (x+1), if you plug in x = oo (oo is the closest thing to the sideways 8 infinity sign on my keyboard) you get oo/oo. But if you take the derivative of the top and bottom, it's 2/1 = 2. The above may not (yet) represent the opinions of my employer. Subject: Re: Could someone explain this... I'm trying to teach myself... I can work derivatives and integrals... but I know that I have to work with limits to understand some of the more complicated theroms. Subject: Plotting Hyperelliptic Curves!!! How do you plot Hyperelliptic Curves? I tried using Maple, Mathematica using Implicit Plot but doesn't give me the exact curve I need. Is there any other way? Thank in advance for any help. Subject: Re: Plotting Hyperelliptic Curves!!! > How do you plot Hyperelliptic Curves? I tried using Maple, Mathematica > using Implicit Plot but doesn't give me the exact curve I need. Is > there any other way? Jacobi functions? what is the function being integrated and its arguments ? Subject: Re: Plotting Hyperelliptic Curves!!! > How do you plot Hyperelliptic Curves? I tried using Maple, Mathematica > using Implicit Plot but doesn't give me the exact curve I need. Is > there any other way? > Thank in advance for any help. > may be you will find this link to a Function renderer useful. It allows you to input any function while producing an excellent graphic output. Remember: it is cardware! (You must send a postcard to the programmer.) Eckard Subject: Re: Cantor Paradox >Nathan's list is the list of all real numbers definable by a finite >string. He must specify definable in what language, say L. >Then the issue is whether the diagonal number for Nathan's list can >be defined in L. Its definition contains such expressions as the >n-th digit of the n-th number defined in L and it is not clear that >these can be defined in L. >Since a contradiction results from the supposition that the diagonal >number can be defined in L, Nathan's paradox actually works as a >reductio ad absurdum to show that it cannot be defined in L. > Well, Nathan evidently thinks it IS clear that the diagonal can be > defined in L. Let L be English. He evidently thinks he has defined > in English exactly how we go about constructing the diagonal. Wow, sounds like Nathan has rediscovered Richard's Paradox (1905). Not bad for an 11-year-old, if he really did think of himself. Subject: Re: Cantor Paradox Discussion, linux) > Wow, sounds like Nathan has rediscovered Richard's Paradox (1905). Not > bad for an 11-year-old, if he really did think of himself. He's not an 11-year-old. Whether he's ever read Richard's Paradox or not, he clearly has the background to choose the issues carefully here. (I hadn't seen the paradox before, so I appreciate having a name attached to it.) -- Contrariwise, continued Tweedledee, if it was so, it might be, and if it were so, it would be; but as it isn't, it ain't. That's logic! -- Lewis Carroll Subject: Re: Cantor Paradox >Nathan's list is the list of all real numbers definable by a finite >string. He must specify definable in what language, say L. >Then the issue is whether the diagonal number for Nathan's list can >be defined in L. Its definition contains such expressions as the >n-th digit of the n-th number defined in L and it is not clear that >these can be defined in L. >Since a contradiction results from the supposition that the diagonal >number can be defined in L, Nathan's paradox actually works as a >reductio ad absurdum to show that it cannot be defined in L. > Well, Nathan evidently thinks it IS clear that the diagonal can be > defined in L. Let L be English. He evidently thinks he has defined > in English exactly how we go about constructing the diagonal. Well, he clearly can't have. It's not particularly clear to me that the n-th digit of the n-th real number defined by a finite string in English is well-defined at all. It's not clear to me that English is that well-defined. If English IS that well-defined, then clearly the n-th digit of the n-th real number defined by a finite string in English can't be a string with a definite meaning in English. The logic is inescapable. > The distinction between paradox and reductio ad absurdum can be > controversial. A mathematical argument reaching an absurd conclusion > calls into question the underlying assumptions of the argument, but > there can be disagreements about which underlying assumption is > faulty. If there is only one possible candidate, the argument is a > reductio ad absurdum. But if it not clear that there is only one > possible candidate then, until we figure things out, we should use > the word paradox. > Nathan evidently does not believe that an ability to describe his > diagonal in L is obviously the sole possible candidate for > faultiness, and so he evidently finds it appropriate to call into > question other propositions upon which his argument is based. He's arguing against Cantor's theorem. That's a theorem of ZFC. If he wants to argue that the propositions on which Cantor's theorem is based lead to contradiction, he should find a contradiction in ZFC. I'm not aware of any definition of English in ZFC. We could let L be the language of second-order arithmetic with abstraction terms. Then the reason Nathan's argument to a contradiction doesn't go through in ZFC is that L can't define the n-th digit of the n-th number defined in L. Nathan hasn't found a contradiction in ZFC. He hasn't given us any reason to doubt Cantor's theorem. > Perhaps Nathan thinks that diagonalization in general is suspect. > Perhaps he thinks there is some problem in the concept of infinity or > of denumerability. Perhaps he thinks there is some much more deeply > rooted proposition whose faultiness he has uncovered. Perhaps he > even thinks there is paradox at the heart of the mathematical > enterprise. Yes, but he hasn't given us any reason to believe that, because he hasn't found a contradiction in ZFC. > Danny Purvis Subject: Re: Cantor Paradox >Even more, Nathans list, if it can exist at all, is already an >infinite string and it must be given before the diagonal can be >constructed. > >Which means that any diagonal requires an infinite string, the >list, for its definition. > I'm having trouble seeing this point. Why would this same objection > not apply to the standard argument showing the reals are not > denumerable? It seems to me that Nathan can construct his diagonal > as far out as he wants to, one step at a time, just the way the > diagonal in the standard argument is constructed, only assuming the > availability of an algorithm for translating strings to real numbers. > Danny Purvis > The diagonal is supposed to, according to Nathan, to require only > finitely many characters for its definition, but the list on which it is > defined must contain infinitely many. If the diagonal cannot exist > without an infinite list, then it requires infinitely many characters > for its definition and cannot be one of Nathan's finite string numbers. > A standard real need not be definable by a finite string of characters, > in fact it is necessary that most of them not be so definable. You are proving your point by assuming what you are trying to prove. We don't have to just consider the diagonal proof. Cantor's first proof, published in 1874, is a finite description of a real number that can not be in the set of all definable numbers. Saying that the diagonal number is not definable is self defeating. If the missing real can't be defined, how can you prove whether it is or is not in Nathan's list? If the missing number is definable, it can't be in Nathan's list, proving the definable reals are uncountable. Russell - 2 many 2 count Subject: Re: Cantor Paradox > > >Even more, Nathans list, if it can exist at all, is already an > >infinite string and it must be given before the diagonal can be > >constructed. > > >Which means that any diagonal requires an infinite string, the > >list, for its definition. > > > I'm having trouble seeing this point. Why would this same objection > not apply to the standard argument showing the reals are not > denumerable? It seems to me that Nathan can construct his diagonal > as far out as he wants to, one step at a time, just the way the > diagonal in the standard argument is constructed, only assuming the > availability of an algorithm for translating strings to real numbers. > > Danny Purvis > > The diagonal is supposed to, according to Nathan, to require only > finitely many characters for its definition, but the list on which it is > defined must contain infinitely many. If the diagonal cannot exist > without an infinite list, then it requires infinitely many characters > for its definition and cannot be one of Nathan's finite string numbers. > A standard real need not be definable by a finite string of characters, > in fact it is necessary that most of them not be so definable. > You are proving your point by assuming what you are trying to prove. > We don't have to just consider the diagonal proof. > Cantor's first proof, published in 1874, is a finite description of a real > number that can not be in the set of all definable numbers. > Saying that the diagonal number is not definable is self defeating. > If the missing real can't be defined, how can you prove whether > it is or is not in Nathan's list? The claim is that the diagonal number is not definable *in L*, where Nathan's list is the list of numbers definable in L. It may be definable in a metalanguage, more powerful than L. We need to use such a language to define Nathan's list in the first place. > If the missing number is definable, it can't be in Nathan's list, > proving the definable reals are uncountable. > > - 2 many 2 count Subject: Re: Cantor Paradox > Saying that the diagonal number is not definable is self defeating. > If the missing real can't be defined, how can you prove whether > it is or is not in Nathan's list? > The claim is that the diagonal number is not definable *in L*, where > Nathan's list is the list of numbers definable in L. > It may be definable in a metalanguage, more powerful than L. > We need to use such a language to define Nathan's list in the first > place. Nathan's set of definable real numbers is just a list of real numbers. We can use any language we want to define these real numbers. Why should we assume Cantor's missing number exists if we can't define this number in some language? I don't see how a metalanguage enters into this. Any language that can be used to define Cantor's missing number can also be used to define Nathan's list. > If the missing number is definable, it can't be in Nathan's list, > proving the definable reals are uncountable. Claiming the diagonal number is only definable in a metalanguage begs the question of whether this missing number even exists. Does this mean Cantor's number is a meta-real number? Russell - the universe is one dimensional Subject: Re: Cantor Paradox > I don't see how a metalanguage enters into this. > Any language that can be used to define Cantor's missing number > can also be used to define Nathan's list. Not necessarily. If it requires a priory existence of an infinite list to define the diagonal, and Nathan's numbers must be definable by finite strings, then that language in which the diagonal is defined CANNOT be used for the numbers in Nathan's list. > If the missing number is definable, it can't be in Nathan's list, > proving the definable reals are uncountable. > Claiming the diagonal number is only definable in a > metalanguage begs the question of whether this > missing number even exists. How so? > Does this mean Cantor's number is a meta-real number? Since Cantor identifies irrational numbers with infinite strings ( a radix point and infinitely many digit characters), none of Nathan's crap is apposite. Subject: Re: Cantor Paradox > Saying that the diagonal number is not definable is self defeating. > If the missing real can't be defined, how can you prove whether > it is or is not in Nathan's list? > > The claim is that the diagonal number is not definable *in L*, where > Nathan's list is the list of numbers definable in L. > It may be definable in a metalanguage, more powerful than L. > We need to use such a language to define Nathan's list in the first > place. > Nathan's set of definable real numbers is just a list of real numbers. > We can use any language we want to define these real numbers. Yes, but if the list is the list of all real numbers definable in a language L, you will have to use a language more powerful than L to define the list, a language in which you can talk about the semantics of L. That's how the metalanguage enters. > Why should we assume Cantor's missing number exists if > we can't define this number in some language? I told you that we could define it in a language, but it can't be L. > I don't see how a metalanguage enters into this. > Any language that can be used to define Cantor's missing number > can also be used to define Nathan's list. > If the missing number is definable, it can't be in Nathan's list, > proving the definable reals are uncountable. > Claiming the diagonal number is only definable in a > metalanguage begs the question of whether this > missing number even exists. It exists as long as the list exists. The list exists as long as the language L has a well-defined semantics. In which case there will be a metalanguage for talking about the semantics of L. > Does this mean Cantor's number is a meta-real number? > > - the universe is one dimensional Subject: Re: Cantor Paradox > Saying that the diagonal number is not definable is self defeating. >> If the missing real can't be defined, how can you prove whether >> it is or is not in Nathan's list? >> >> The claim is that the diagonal number is not definable *in L*, where >> Nathan's list is the list of numbers definable in L. >> It may be definable in a metalanguage, more powerful than L. >> We need to use such a language to define Nathan's list in the first >> place. > Nathan's set of definable real numbers is just a list of real > numbers. We can use any language we want to define these real > numbers. Why should we assume Cantor's missing number exists if > we can't define this number in some language? Nathan's set is the set { x in R | exists formula P such that |= P(x) & E!y P(y) } How do you define this set in the object language without being able to represent |=? Russell, I'm not trying to be insulting here, but if you don't understand that last question, you're not likely to understand the issues here. If you want to pursue this conversation and you really don't know what I mean by that last question, then you'll need some background reading. > I don't see how a metalanguage enters into this. > Any language that can be used to define Cantor's missing number > can also be used to define Nathan's list. >> If the missing number is definable, it can't be in Nathan's list, >> proving the definable reals are uncountable. > Claiming the diagonal number is only definable in a > metalanguage begs the question of whether this > missing number even exists. > Does this mean Cantor's number is a meta-real number? This nonsense question confirms that you simply don't understand object language/metalanguage issues. (Please note: I only claim a superficial understanding myself.) -- Customers have come to SCO asking what they can do to respect and help protect the rights of the SCO intellectual property in Linux. SCO has created the Intellectual Property License for Linux in response to these customers needs. -- SCO responds to needs. Subject: Re: Cantor Paradox > > >Even more, Nathans list, if it can exist at all, is already an > >infinite string and it must be given before the diagonal can be > >constructed. > > >Which means that any diagonal requires an infinite string, the > >list, for its definition. > > > I'm having trouble seeing this point. Why would this same objection > not apply to the standard argument showing the reals are not > denumerable? It seems to me that Nathan can construct his diagonal > as far out as he wants to, one step at a time, just the way the > diagonal in the standard argument is constructed, only assuming the > availability of an algorithm for translating strings to real numbers. > > Danny Purvis > > The diagonal is supposed to, according to Nathan, to require only > finitely many characters for its definition, but the list on which it is > defined must contain infinitely many. If the diagonal cannot exist > without an infinite list, then it requires infinitely many characters > for its definition and cannot be one of Nathan's finite string numbers. > A standard real need not be definable by a finite string of characters, > in fact it is necessary that most of them not be so definable. > You are proving your point by assuming what you are trying to prove. > We don't have to just consider the diagonal proof. > Cantor's first proof, published in 1874, is a finite description of a real > number that can not be in the set of all definable numbers. Cantor's first proof does not contain a definition of such a number but merely a proof that at least one such number has to exist if the reals are to have the LUB property. Subject: Re: area of a circle In-reply-to: Ignacio Larrosa Ca.96estro En el mensaje:404F0FEA.8010603@free.invalid, >philippe 92 escribi.97: >> [...] [Pizza theorem] >>> I came up with a pretty short proof based on the fact that when you >>> have two intersecting chords, opposing arcs on the circle sum to >>> twice the angle between the chords. Since the chords are >>> perpindicular, we have that the opposing arcs must be supplementary. >>> Sliding these arcs and chords together, we see that the chords form >>> a right triangle with the diameter of the circle as the hypotenuse. >>> See >>> http://www.whim.org/nebula/math/images/perpchord.gif >> [...] >> Fine ! >> I had forgotten this general property of two intersecting chords. >> This improves the Pizza Theorem : >> Not only the areas are equal, but also the portions of the edge (where >> usually there is no toping...). >Certainly, Rob Johnson's proof is much more easy and pretty. >That didn't seems true for me is that is said in MathWorld: >If a circular pizza is divided into 8, 12, 16, ...slices by making cuts at >equal angles from an arbitrary point, then the sums of the areas of >alternate slices are equal. >(http://mathworld.wolfram.com/PizzaTheorem.html) >If the number n of slices is 4 multiple, but not 8 multiple, it is not >true, or at least, it isn't supported by the mentionated proofs. If n = 12, >by example, you can get 3 bunchs of 4 slices with equal sum of areas, but >not 2 bunch of six alternate slices verifying that. The correct version of the theorem is givent in paper57.pdf. You are correct that the version given in PizzaTheorem.html on MathWorld is not correct for 8n+4 pieces. For 4n pieces and n people, each person needs to take every nth piece, thereby getting 4 pieces each and getting the same amount of pizza. >your first message, >http://www.maths.unsw.edu.au/~mikeh/webpapers/paper57.pdf. >(the 'crust' in the corollary) Rob Johnson Find a configuration of five circles, where each pair > intersects under the same angle @ > (cos @12=(r1^2+r2^2-d12^2)/(2r1r2) - complex @ are > allowed, r=0 or oo (points and lines) too.) > * What is the angle between a point and a circle? (0 degrees, 180 degrees?) > * What would it be good for to allow lines? - > Can't we transform lines into circles > by application of an appropriate angle-preserving map? http://www.channel1.com/users/bobwb/synergetics/photos/ x6girdle6.html --Give Earth a Trickier Dick Cheeny -- out of office, after GIGA years. http://www.benfranklinbooks.com/ http://www.rand.org/publications/randreview/issues/rr.12.00/ http://members.tripod.com/~american_almanac Subject: Re: 5 circles schrieb Brian Quincy Hutchings : > sounds like Leibniz's *pentagrammum mirificum*, > or what ever it's called, five geodesics on sphere. Are you thinking of Gauss's Pentagramma Mirificum??? http://134.76.163.65/servlet/digbib?template=view.html&id= 136942&startpage=1 12&endpage=124&image-path=http://134.76.176.141/cgi-bin/ letgifsfly.cgi&image- subpath=/3580&image-subpath=3580&pagenumber=112&imageset-id= 3580 But this pentagram consists of five circles (great circles even), for which the i-th circle intersects circles (i-1) and (i+1) *only* at a 90 degree angle (for i=1,2,3,4,5), but circles 1 and 3, for example, do meet at a different angle. Am I missing something? - I thought, Hauke asked for five circles such that *all* of the 20 pairwise intersections angles were equal!? [ ... ] > I didn't get the notation in the equation; cosine of alpha-sub(1,2) equals radius-sub(1) to the second power.... d-sub(1,2) is the distance between the centers? Looks O.K. - Given two circles with midpoints M1, M2 and radii r1, r2 which intersect at a point I under an angle alpha, the distance from I to M1 is r1, the distance from I to M2 is r2, and the angle at I is either alpha or 180 degree minus alpha (depending on how you define the intersection angle of two circles), so that by the theorem of cosines the square of the distance between M1 and M2 is given by d12^2 = r1^2+r2^2- 2*r1*r2*cos(alpha) >> Find a configuration of five circles, where each pair >> intersects under the same angle @ >> (cos @12=(r1^2+r2^2-d12^2)/(2r1r2) - complex @ are >> allowed, r=0 or oo (points and lines) too.) Sorry, but the links you appended were not particularly useful to me. Thomas --- Ich hab Dich leider nicht verstanden, aber sag das bitte nicht nochmal. Heinz Rudolph Kunze Subject: Re: 5 circles I formulated it this way because the usual determinant approach gives the unique value @=120 deg...but you can't draw it, some circles will always have radii being NOT positive real. -- Hauke Reddmann <:-EX8 fc3a501@uni-hamburg.de als man ankam wollte man werden, die geschichte schreiben, die doofen sollen sterben, der plan als man damals nach hamburg kam (Kettcar) Subject: Re: 5 circles > I formulated it this way because the usual determinant > approach gives the unique value @=120 deg...but you > can't draw it, some circles will always have radii > being NOT positive real. I see! - So you were using distance geometric methods, too. Are you interested in the Maple worksheet with the calculations? Subject: Re: The Dirac Delta? >>(It's actually the integral of the function 1 with respect to the >>delta function which equals 1. Saying that the integral of >>the delta function is 1 would imply that it's Lebesgue integrable, >>which it certainly is not.) > Bing! I was waiting for someone to say this. > I remember my first real QM class, and the dirac function was > introduced. It was the first time I had to seriously work on > understanding it. Finally the instructor understood what I wanted in > my persistent questions in his office (not to challenge the physics, > but to understand how to make the math rigorous) and he said we never > use the delta function except to multiply it with some other function > and take an integral. Still not rigorous, but with that proviso my > brain clicked. > Once more, I do wish physicists would stop using physicist > pseudo-math. But still, there are many things I wish for, and this is > hardly the only one I won't get. > Can we agree to stop calling the dirac delta function a function > now? > Thomas Dirac's Delta is a functional (member of the dual space of a proper function space - e.g. infinitely derivable function with compact support). This function space is isomorphic to a sub space of the dual space. Some of the results in the dual space have ready match in the function space, but some no. E.g., heaviside function is a real function. Dirac's delta is not (look at the common physicist definition - zero all over except in the origin with integral one), just work out lebesgue integral from the definition. However, other concepts are transferable to the real functions - e.g., convolution with the delta function is defined. This is similar situation to extending the real field into the complex one, working some stuff there, and getting back results for the reals. Try any book on functional analysis (the classic one being by Dunford & Schwartz). Subject: generalization of stokes' and divergence theorems? Stokes theorem says SS[ curl F dS ] over S = S[ F dr ] over C, and the divergence theorem says SSS[ div F dV] over E = SS[ F dS ] over S. Is there a theorem that says, SSSS[ something(F) dH ] over H = SSS[ F dV] over V. Wouldn't V have to have an orientation? It seems that in R^3 we couldn't have that, but in R^4 I'd think you could define something like that. Is there a general something where I can write: SS...n[ something(F) dA ] over A = SS...(n-1)[ F dB] over B or something similar? Thanks, Jeremy Subject: Re: generalization of stokes' and divergence theorems? > Stokes theorem says SS[ curl F dS ] over S = S[ F dr ] over C, and the > divergence theorem says SSS[ div F dV] over E = SS[ F dS ] over S. > Is there a theorem that says, SSSS[ something(F) dH ] over H = SSS[ F dV] > over V. Wouldn't V have to have an orientation? It seems that in R^3 we > couldn't have that, but in R^4 I'd think you could define something like > that. > Is there a general something where I can write: > SS...n[ something(F) dA ] over A = SS...(n-1)[ F dB] over B > or something similar? http://en.wikipedia.org/wiki/Stokes'_theorem -- P.A.C. Smith The vast majority of Iraqis want to live in a peaceful, free world. And we will find these people and we will bring them to justice. Subject: Re: generalization of stokes' and divergence theorems? The theorems you cite (Green, Stikes, Grad, Div, Curl, etc. are from vector field theory in 3d. The ones that you wish to obtain require an understanding of differential geometry on arbitrary manifolds. Check out Michael Spivak's work or that of Harley Flanders Subject: Re: generalization of stokes' and divergence theorems? > The theorems you cite (Green, Stikes, Grad, Div, Curl, etc. are from > vector > field theory in 3d. The ones that you wish to obtain require an > understanding of differential geometry on arbitrary manifolds. The term differential *geometry* could be misleading; perhaps calculas on manifolds fits best. -- I'm on warm milk and laxatives Cherry-flavored antacids reverse my forename for mail! - saibot Subject: Re: generalization of stokes' and divergence theorems? >Check out Michael Spivak's work or that of Harley Flanders I liked Spivak's _Calculus on Manifolds_. Keith Ramsay Subject: Help with equation Can anyone tell me what the mathematical equation is for the following statement: The whole is worth more than the sum of its parts. Thanks in advance, xillion Subject: Re: Help with equation In sci.math, xillion : > Can anyone tell me what the mathematical equation is > for the following statement: > The whole is worth more than the sum of its parts. > Thanks in advance, > xillion The equation? This is far too vague a specification. However, one could be slightly arbitrary; the following equation is a more or less literal transilation of the statement: S = sum(i,+oo) s_i > S_n = sum(i=1,n) s_i (in more traditional math texts 'sum' would be a large sigma). and this is true for any convergent series where each term is positive. For example, S = sum(i=1,+oo) (2^(-i)) = 1 and S_n = sum(i=1,n) (2^(-i)) = 1 - 2^(-n) as can easily be verified. But this is one of an infinite number of examples; another one, for instance, is e^x = sum(i=0,+oo) x^i/i! > sum(i=0,n) x^i/i! for x > 0. Another example, which does not converge, is +oo = sum(i=1,+oo) 1/i > sum(i=1,n) 1/i Unfortunately, the whole is not necessarily an infinite series, its parts is not necessarily its terms, and worth does not refer to numeric value. As a philosophical statement, it's quite profound. As a mathematical statement, it's far too vague. -- #191, ewill3@earthlink.net It's still legal to go .sigless. Subject: Re: Help with equation Gestault=True > Can anyone tell me what the mathematical equation is for the following statement: > The whole is worth more than the sum of its parts. > Thanks in advance, > xillion Subject: Re: Help with equation > Can anyone tell me what the mathematical equation is for the following > statement: > The whole is worth more than the sum of its parts. > Thanks in advance, > xillion What makes you think there is one? Subject: Re: Random pi digits and Mahler's theorem charset=Windows-1252 [typo correction] Extending the previous results (well-known, no doubt) ... Theorem. Let X be uniformly distributed on interval (a, b), where a and b are integers, and let m be a real number > 1. Then pr( exist integers p,q > 1, |X - p/q| <= 1/q^m ) <= 2(Zeta(m)-1) where Zeta is the Riemann Zeta function. Proof (sketch): pr( exist integers p,q > 1, |X - p/q| <= 1/q^m ) <= sum( pr(|X - p/q| <= 1/q^m, integers p,q > 1 ) (subadditivity) <= sum( 2(1/q^m)/(b-a) + (q(b-a)-1)(2/q^m)/(b-a), q > 1) [*] <= sum( 2/q^(m-1), q > 1 ) <= 2 (Zeta(m-1) - 1) where [*] follows by noting the cases summarized in the following table, each case corresponding to the joint occurrence of the conditions indicated in the first two columns of a row: contrib. to sum p/q - 1/q^m p/q + 1/q^m cases #cases (per case) -------------------------------------------------------------- --- <= a > a p = aq 1 (1/q^m)/(b-a) < b >= b p = bq 1 (1/q^m)/(b-a) > a < b aq < p < bq q(b-a)-1 (2/q^m)/(b-a) other other other 0 -------------------------------------------------------------- --- QED. Some examples: m 2 ( Zeta(m-1) - 1) ----------------------- 4 0.40411... 5 0.16464... 6 0.73855... 10^-1 7 0.34686... 10^-1 8 0.16698... 10^-1 9 0.81547... 10^-2 10 0.40167... 10^-2 15 0.12249... 10^-3 20 0.38164... 10^-5 30 0.37253... 10^-8 40 0.36379... 10^-11 42 0.90949... 10^-12 --r.e.s. Subject: Re: Random pi digits and Mahler's theorem charset=Windows-1252 Extending the previous results (well-known, no doubt) ... Theorem. Let X be uniformly distributed on interval (a, b), where a and b are integers, and let m be a real number > 1. Then pr( exist integers p,q > 1, |X - p/q| <= 1/q^m ) <= 2(Zeta(m)-1) where Zeta is the Riemann Zeta function. Proof (sketch): pr( exist integers p,q > 1, |X - p/q| <= 1/q^m ) <= sum( pr(|X - p/q| <= 1/q^m, integers p,q > 1 ) (subadditivity) <= sum( 2(1/q^m)/(b-a) + (q(b-a)-1)(2/q^m)/(b-a), q > 1) [*] <= sum( 2/q^(m-1), q > 1 ) <= 2 (Zeta(m) - 1) where [*] follows by noting the cases summarized in the following table, each case corresponding to the joint occurrence of the conditions indicated in the first two columns of a row: contrib. to sum p/q - 1/q^m p/q + 1/q^m cases #cases (per case) -------------------------------------------------------------- --- <= a > a p = aq 1 (1/q^m)/(b-a) < b >= b p = bq 1 (1/q^m)/(b-a) > a < b aq < p < bq q(b-a)-1 (2/q^m)/(b-a) other other other 0 -------------------------------------------------------------- --- QED. Some examples: m 2 ( Zeta(m) - 1 ) ---------------------- 4 0.40411... 5 0.16464... 6 0.73855... 10^-1 7 0.34686... 10^-1 8 0.16698... 10^-1 9 0.81547... 10^-2 10 0.40167... 10^-2 15 0.12249... 10^-3 20 0.38164... 10^-5 30 0.37253... 10^-8 40 0.36379... 10^-11 42 0.90949... 10^-12 --r.e.s. Subject: (clarification) chapter 2 in math faq There have been many threads on are the natural numbers in the integers... and I understand the usual responses to these threads (essentially what you said below). But my original question was about the domains of the operations on the sets Z,Q,... , not the sets. What type of ignoring goes on here? Is + just domain-less when it doesn't matter? Ie, in elementary number theory, frequently there is division, multiplication by sqrt(2), etc. Obviously we can't always work over only (Z,+_Z,*_Z). Is it understood that there is a constant switching of number system _operations_, or something else? Thank you, Khintchine > the math faq talks about construction of the number systems. in 2.2.8 > it mentions that the naturals are not a subset of the integers, etc. it > defines some embeddings so they can be considered to be subsets. > so it's unambiguous to talk about the number 3. what about the expression do people generally imagine little Z,Q,R,C's floating around under the + > signs? Fishfry said, > No, people just ignore these fine points and imagine the integers being > a subset of the rationals being a subset of the reals. The only time the > natural embeddings are important is if you are discussing the > construction of the number systems. Subject: stochastic group S(2,4) How to prove it is isomorphic to alternating group of order 12? Subject: Re: stochastic group S(2,4) > How to prove it is isomorphic to alternating group of order 12? Um, what is it? -- Robin Chapman, www.maths.ex.ac.uk/~rjc/rjc.html Lacan, Jacques, 79, 91-92; mistakes his penis for a square root, 88-9 Francis Wheen, _How Mumbo-Jumbo Conquered the World_ Subject: Re: stochastic group S(2,4) >> How to prove it is isomorphic to alternating group of order 12? >Um, what is it? Never heard of it. S(2,4) is sometimes used for symplectic group Sp(2,4), but that is the same as SL(2,4) and is isomorphic to alternating group of order 60. It's not SO^+(2,4) or SO^-(2,4) either - they have orders 6 and 10. Derek Holt. Subject: Re: Proofs >I would like to know how I may approach if someone asks me to prove a >wrong statement. For example if someone asks me the statement 'prove >that 101 is an even integer'. > In this case I would suggest that you prove that the statement is false. > In your case, this is rather obvious, e.g. you could assume that there > is an integer k with 2*k=101, but then 50 Anyway, this statement is obviously false, so you could just say that > and that's all. > If someone asked me to prove that 101 is an even integer, I would first > mention to that person that he must not be dealing with decimal numbers, > because in base ten, 101 is not even. Then I would prove to him that 101 > in an odd base is even. Indicating that context is everything. This could be a quite serious question on a Friday afternoon (especially at Coslo's or whatever the local equivalent is), or in a problem solving class, or just somewhere where clear statement of assumptions is important (or even critical) (are you listening, NASA: a billion dollars here, a billion dollars there, and pretty soon you're talking real money). Jon Miller Subject: Re: tensors >I'm understanding tensors are things who obey certain transformation >properties and algebraic properties and everything, but I don't understand >what they physically/geometrically mean. Have a look at http://math.ucr.edu/home/baez/gr/gr.html Keith Ramsay Subject: Re: tensors > I'm understanding tensors are things who obey certain transformation > properties and algebraic properties and everything, but I don't understand > what they physically/geometrically mean. What is the physical significance > of a covariant, contravariant, or mixed tensor of a specific order? I.e., > what's the physical difference between A^ij, A_ij, and A^i_j? Would 'columns of columns', 'rows of rows', and 'rows of columns' satify the question or are you looking for some physical orientation explanation like, whats the significant diferrence between 'left handed' and 'right handed' viewpoints? There's a prevailing viewpoint under which 'column vectors' transform covariantly, but if you find linearly independent functionals in that space, they form a dual space of contravariant dual row vectors. The two spaces are images of the same 'physcal' phenomenon, its just convention that begins with seeing it one way first. > I'm also > confused as to why we write a dot product of vectors in tensor notation as > a_i b^i, why not a_i b_i, or a^i b^i? Is there a difference between a^i b_i > and a_i b^i? I guess I'm not really understanding the difference between > covariant, contravariant, and mixed, other than that the differences in how > they transform. This is just for consistency. Group operations are dependent on left or right action. Those transforms don't let you jiggle them around willynilly Say for a 2x3 matrix, the row dimension is 2 and column dimension is 3, and that limits what you can do with the rows versus column operations. Hope this helped a bit. MK Subject: Re: tensors Schaum's Outline on Vector Vnalysis will help you keep track of your Christoffel Symbols when understanding and manipulating tensors. A good book on Differential Geometry will help. Try Dirk Struik (MIT), Dover Publ. Subject: Re: simulating bivariate exponential >I am looking for a way of simulating from the bivariate exponential distribution given by Marshall & Olkin: >F(x,y) = exp{-[Micro]1x-[Micro]2y-([Micro]3+d)max(x,y)} >in order to simulate a trial with surrogate and primary endpoint failure times. Is that the tail distibution (P{X>x, Y>x})? The unnormalized density? It certainly cannot be the CDF. How about generating one variate according to the marginal and the other according to the conditional distribution? -- Stephen J. Herschkorn herschko@rutcor.rutgers.edu Subject: Re: Math Shirts (again) > Hey, I just wanted to update (see: > http://groups.google.com/groups?dq=&start=25&hl=en&lr=&ie=UTF- 8&oe=UTF-8&gro > up=sci.math&selm=4bc08be9.0312282107.564f1695% 40posting.google.com > ) > Got a new url going, now it's simply http://www.mathshirts.com and I'm > adding new designs as I think of them > Comments, critiques, congratulations and cognac are all welcomed and > invited. > I have seen studnets wearing shirts bearing Maxwell's equaions. Maybe > something of that level of tech-geeky-inside-jokiness would be more > attractive. Yeah, I've seen them too .. and that's the problem, I don't want to just make designs that are already out there, if you want that, go for it, but in all honesty, I'm making these designs for myself, and I figure there are going to be some people out there with similar tastes. I feel like there's nothing besides inside jokes, equations on shirts and elementary school teacher designs. Subject: Re: Math Shirts (again) > I have seen students wearing shirts bearing Maxwell's equations. May be > something of that level of tech-geeky-inside-jokiness would be more > attractive. It is nevertheless likeable, this one may be ? http://www.scienceteecher.com/astronomy_all.htm Mental images have spurred many a mathematician, physicist and chemist during their creative persuits.... Among the equations and images I consider T-Shirt display worthy : and the negatively curved vortex as model for hyperbolic geometry. F= m a and the falling apple and the rocket. e ^ [ 2 Pi n i] = 1 Euler's equation Euler-Bernoulli Equation of Variational calculus _ Action Integral: Max _/ (T-V)dt D N A molecule with cross bands of Crick/Watson Zhukowski Airfoil, W= z+ a^2/z, and the Kitty Hawk WB aircraft Fractal images Mandelbrot / Julia anmd FeigenBaum diagram Platonic Solids and the mixed surface Icosidodecahedron with its great circles. Pyramids of Egypt. Schroedinger's Equation Subject: Re: David Ullrich's parachute <$D$zGFA64NUAFwMT@meden.demon.co.uk Discussion, linux) >>> Working quickly, Mr. Ullrich used his sand-making device to add >>> an infinite amount of new sand to his backpack. At infinity, the >>> Earthward forces and the Moonward forces balanced eachother perfectly, >>> and our friend remained suspended in mid-air until the Air Force could >>> rescue him. >>I'm not a physicist, so you'll have to help me here. Does he achieve >>equilibrium before or after the black hole forms? > After I think, as the perfect equilibrium is approached as the mass > tends to infinity. > However, he may be hit by the Earth, moving at high velocity, before the > black hole is formed, but this depends on exactly how fast the sand is > produced. > But, if the infinite amount of sand is produced instantaneously > equilibrium is obtained and the black hole created at the same time. I think that Nathan has taught us an important lesson. If we allow infinite mass in a very limited space in our physical models (and also decide to suspend the conservation of matter/energy), then counterintuitive things happen. Obviously, we must reject Cantor. (Ignoring, for the moment, the idea of limits tending to infinity predating Cantor and the means of comparing two such limits by their limit of their ratios.) Nathan, not your best effort. You're trying to hard. Go play soccer and take your mind off it for an afternoon. -- My proofs are out there. -- James S. Harris Subject: need help on prob and stat~ Thanks 1) There are k taxis in Vancouver, numbered consecutively 1 to k. Suppose that the taxi that brings you from the airport is number 123. (a) Devise a suitable model for this situation. Give an unbiased estimate for k. (b) A friend claims to have heard that k 1000. Do the observed data provide grounds to reject the claim at the 5% signi cance level? (Hint: under the claim, what is the probability your taxi number would be this small?) (c) Suppose your next taxi is number 200. Answer (a) and (b) again using all the data now available. (Hint: in (a), consider the sum of the two taxi numbers). 2) Let X1,X2.... be iid with mean and variance. Let Sn^2 be the sample variance of X1,...Xn. Show lim n-> infinity Sn^2= var with probability 1. Subject: Gaston Julia's Sets I took a the escape theory of the series for the Gaston Julia's set and remind the reader to learn why it is an escape number to define the set. Gaston's Differentail- defined using the escape number from a Gaston Julia's Set website ************************************************************** ****** The pioneering work of the French mathematician Gaston Julia, published in 1918 when Julia was 25 years old, was essentially forgotten by the mathematical world until the 1980's, when computers made possible the visualization of his creation. Julia's idea was to observe the behavior of the orbit of a complex number (see Vignette 6 and Vignette 10) under iteration of a function f . That is, begin with a complex number z0 , visualized as a point in the plane, and apply f to z0 . The resulting value is fed back into the function f to obtain a new complex number z1. This in turn is fed back into f to obtain z2, and so on. The resulting sequence of complex numbers {z0 , z1 , z2 , ...} is called the orbit of z0 under f . As an example, consider the function . Starting with z0 = 0, we obtain the following orbit: The terms in this sequence remain bounded -- that is, they never get very far from 0, the origin of the complex plane. On the other hand, for the same function f, the orbits of other starting points may move arbitrarily far away from 0. As an example, if we start with z0 = 2, we get the following orbit: We will refer to a complex number z0 as a prisoner if its orbit under f is bounded, and an escapee if the orbit is unbounded -- that is, terms in the orbit get arbitrarily far away from 0. The set of all prisoners for a given function f is called its prisoner set, and the set of all escapees is called the escape set. In the above examples, we used complex numbers z0 for which the imaginary part is 0, just to simplify the calculations. But the analogous calculations can be done for any complex value of z0 , using the multiplication and addition of complex numbers as discussed in Vignette 10. For example, for the function , the orbit of is In this example, we see that is in the prisoner set for , because its orbit is bounded. Julia Sets Julia was interested in the properties, for various functions, of the prisoner set and the escape set -- and also what is now called the Julia set for the function. The Julia set is defined to be the boundary between the prisoner set and the escape set. The Julia set for f consists of the points In order to determine these sets experimentally, it is necessary to compute the orbits of a great many starting points z0 , since we must know the orbit in order to determine whether the starting point is a prisoner or an escapee. And if we had to do the calculation of orbits such as the last one above purely by hand -- for several thousand starting values -- we would quickly lose interest in the project! That's where mathematical theory comes into play. Even without actually doing all of the tedious calculations, a number of things can be proved about prisoner sets, escape sets and Julia sets. But rather than looking at the theory, we will instead consider the fact that, starting in the 1980's, the availability of computers led people back to the idea of a visual exploration of Julia sets -- with incredibly beautiful results. After all, a computer could be programmed to do the calculation of an orbit of a starting value of z0 -- and would not even object if asked to do a similar calculation for thousands or even millions of such similar values! For simplicity, let's consider only functions of the form , where c is some fixed complex number. It can be proven that if the orbit of a starting value z0 ever leaves the circle of radius 2 centered at the origin, then the rest of the terms in the orbit get successively farther from 0, and in fact the orbit is unbounded. Such a starting value z0 would therefore be an element of the escape set for f . Armed with this simple mathematical theorem and a good bit of computer power, we are ready to visualize the prisoner set and escape set for a given function of the form . The Julia set can then be pictured as the boundary between the two sets. ************************************************************** ************* Subject: Re: Gaston Julia's Sets > Julia was interested in the properties, for various functions, of the > prisoner set and the escape set -- and also what is now called the > Julia set for the function. The Julia set is defined to be the > boundary between the prisoner set and the escape set. These are oversimplifications intended for non-mathematicians. For correct versions, see (for example) Devaney's book -- G. A. Edgar http://www.math.ohio-state.edu/~edgar/ Subject: Re: question about topology Rup > Sorry, why does H_k(R^n,R^n{x})=H_k(S_n-1)? H_k(R^n,R^n{x})=H_k(Rn{x}) Just another excison. (with a little unimportant problem for k=0 I don't want to discuss) R^n{x} deformation retract to S_n-1. So they have same homotopy type and same homology type. It is this way of thinking that make homotopy and holology so powerfull. If you are interested in the topics. Read the chapter about homology in Hatcher's book Algebraic Topology. Just a few hours to understand the stuff. Other chapters or other books like the great book of Spanier are more difficult. It can be foung online. Subject: New source of the math!!! I have come across a source providing good support to those experiencing difficulties with doing homework in maths including arithmetic(http://www.bymath.com/studyguide/ari/form1.htm), geometry (http://www.bymath.com/studyguide/geo/pro/pro1/pro1.htm), algebra (http://www.bymath.com/studyguide/alg/pro/pro1/pro1.htm), functions and graphics (http://www.bymath.com/studyguide/fun/pro/pro.htm), principles of analysis(http://www.bymath.com/studyguide/ana/pro/pro1/pro1. htm), all this is also supported by a good deal of examples and illustrations. Subject: Re: Die Petry, die >> It clearly makes sense to talk about >> natural numbers in relation to computation. Sets of natural >> numbers, sets of sets of natural numbers, sets of sets of sets >> of natural numbers and so on... these are the familiar kinds >> of sets by which set theorists model most ordinary mathematics. >> At what point does it become meaningless? I don't see that >> it does. Is the complaint only about higher set theory? > Over the past 15 years, in both newsgroup discussions and > in private email discussions, I have expended a great deal > of energy trying to explain to you the idea of observability > and the motivations for it. It is clear to me that all that > energy was wasted. Yes. -- Robin Chapman, www.maths.ex.ac.uk/~rjc/rjc.html Lacan, Jacques, 79, 91-92; mistakes his penis for a square root, 88-9 Francis Wheen, _How Mumbo-Jumbo Conquered the World_ Subject: Re: Trapezoidal Rule & Simpson's Rule - Comparison > I'm looking for a simple and obvious example where the Trapezoidal Rule > would give a better approximation to a given definite integral than > Simpson's Rule. Assume the number of subintervals used is the same for >both approximations. For f:[A,B]-->R integrable on [A,B] let us denote I(f)-INTEGRAL_{x=a to x=b}f(x) dx , -infty < A =< a < b =< B < infty , Consider the quadrature formulas b-a (1) I(f)= ----- ( f(a)+ f(b) ) + RT(f) 2 b-a a+b (2) I(f)= -----( f(a) +4*f(-----) + f(b) ) +RS(f) 6 2 where RT(f)=RT(f;a,b) , RS(f)=RS(f;a,b) are the remainders in trapezoidal formula respectively in Simpson (Kepler) quadrature. Observe that 2*(b-a) f(a)+f(b) a+b RT(f)-RS(f)= -------( --------- - f(-----) ) . 3 2 2 Therefore we have following assertion : Proposition. The inequality ============ RT(f;a,b) <= RS(f;a,b) for all [a,b] , subset in [A,B] , if and only if f:[A,B]---> is a concave function. ===== In fact, If f is in C^{2}[a,b], m => I'm looking for a simple and obvious example where the Trapezoidal Rule >> would give a better approximation to a given definite integral than >> Simpson's Rule. Assume the number of subintervals used is the same for both >> approximations. >Let f(x) = x^(2/5) on the interval [-1,1] and use one subinterval. I assume you mean two subintervals, i.e. you evaluate f at the three points -1, 0 and 1, with Trap = 1/2 f(-1) + f(0) + 1/2 f(1) and Simpson = 1/3 f(-1) + 4/3 f(0) + 1/3 f(1). A singularity is not necessary, of course: try cos(pi x) which the trapezoidal rule (for these points) happens to get exactly right but Simpson doesn't. And this generalizes easily to give examples with more subintervals. Robert Israel israel@math.ubc.ca Department of Mathematics http://www.math.ubc.ca/~israel University of British Columbia Vancouver, BC, Canada V6T 1Z2 Subject: re:Die Cantor Die Godel gave a big blow to all those who advocated Petry's views at his time. Of course the examples of undecidable statements he gave aren't very important in applications (unless you need the consistency of arithmetics to sleep at night), but later more natural examples have been found, some of them about simple games like the hydra game. Few believe now that the Riemann Hypothesis is unprovable in ZF, but is it provable in finite arithmetics? It sure is useful to derive properties of finite arithmetics that are important in applications. Some results in operator theory that motivated some quantum mechanical models probably lie outside the scope of finite arithmetics too. Examples of genuinely undecidable statements of finite arithmetics are found all the time, but more importantly, many useful results are obtained from ZF that no one knows how to prove (or sometimes even formulate) by finite methods. Petry's program would require that we abandon the concept of proof and focus on just using the results we believe to be true. But historically it was always the search for proves that driven mathematical investigations. So the results will cease arriving and we'll have no new mathematics to apply. Posted Via Usenet.com Premium Usenet Newsgroup Services ---------------------------------------------------------- ** SPEED ** RETENTION ** COMPLETION ** ANONYMITY ** ---------------------------------------------------------- http://www.usenet.com Subject: Re: Die Cantor Die > And, to repeat myself because so many of my detractors > never seem to catch on, formal systems themselves are > objects that live in the world of mathematics as I > envision it. Except that calculations in them are arbitrarily excluded from Petry's subjective objective world of computation. -- Robin Chapman, www.maths.ex.ac.uk/~rjc/rjc.html Lacan, Jacques, 79, 91-92; mistakes his penis for a square root, 88-9 Francis Wheen, _How Mumbo-Jumbo Conquered the World_ Subject: Re: Die Cantor Die > The basics ideas I stress are: (1) mathematics can be > treated as a science Well there we are. That's a fundamental blunder. -- Robin Chapman, www.maths.ex.ac.uk/~rjc/rjc.html Lacan, Jacques, 79, 91-92; mistakes his penis for a square root, 88-9 Francis Wheen, _How Mumbo-Jumbo Conquered the World_ Subject: Re: Die Cantor Die > We don't lose any of our capacity to understand > the reality we live in if we believe that the purpose of all > abstract thinking is to relate the phenomena we observe in > the real world to phenomena we observe in the world of > computation. > Why would I think the purpose of all abstract thinking is to relate > observable phenomena to computational phenomena? > The idea is that the mind can be understood by analogy with > artificial intelligence software running on a computer (the > brain). Shorly after electricity was discovered many people thought the mind could be understood as electrical interactions. Frog legs could be made to twitch with a spark, etc. Nowadays, we know that thinking does involve electrical impulses, but few people would say that the mind can be understood by anology with electricity. Every generation thinks thinking can be explained by the latest cool idea. Certainly, the mind does have a computational element, but there is a lot more to the human mind than computation. I know. I've written AI programs and they don't think like humans. This is probably a good thing. I wouldn't want a human trying to control a large chemical plant or trying to land the Space Shuttle from orbit. Russell - Zeno was right. Motion is impossible. Subject: Re: Die Cantor Die > Certainly, the mind does have a computational > element, but there is a lot more to the human mind than > computation. Well, Penrose would agree with you. But I've never seen an argument for your view that I find remotely convincing. Presumably the workings of the brain obey the laws of quantum mechanics, which in turn could be modelled in a large computer. So maybe you're saying the mind is more than the workings of the brain. I will not continue this line of discussion. Subject: Re: Die Cantor Die > Mathematics has ALWAYS been an exquisite game, with its own > rules and in its own world. Wanna play? Accept the rules. > In deed, that's seems to be the way the pure mathematicians > view it. To everyone else - those who actually apply it - > mathematics is more than just a game. There is an underlying > reality to it, which is computation. > It's not the computation, it's the utility, that everyone else is > interested in. > But without the pure mathematicians to create the stuff that these > non-mathematicians are so eager to profit from, it wouldn't be there for > them at all. Many advances in pure mathematics were motivated by practical concerns. The Egyptians figured out how to take the roots of numbers so they could build better pyramids. Newton developed calculus to describe the motion of planets. I think pure and applied mathematics complement each other. Russell - 2 many 2 count Subject: Re: Die Cantor Die > I think pure and applied mathematics complement each other. > > - 2 many 2 count And as long as neither tries to inhibit what the other does, it shall remain so. Subject: Re: Die Cantor Die >>Which of the axioms do you object to? > The objection is to the logic as much as the axioms. My first reaction on reading this statement is to suspect that you have a disagreement with the workings of formal mathematical logic. If so, you may find that you have a lot of work to do to get anyone to join you in changing the foundations of mathematics. If I am misunderstanding your meaning, then I have no clue what you are trying to say. > Certainly the usual constructivist objections to the law > of the excluded middle and the axiom of choice are > things that I accept but do not stress. (They are > not fundamental objections) Ok. > The basics ideas I stress are: (1) mathematics can be > treated as a science which requires statements to > have falsifiable implications Falsifiable against what? The best you can do is say that a certain section of mathematics is not a model of reality. I view it as the job of physicists, chemists, etc to determine which mathematical models best represent reality. > (2)when we observe > infinitary objects, we are observing approximations, > and the logic we use must then deal with uncertainty There is already a branch of mathematics that deals with these uncertainties without the need to change the foundations of mathematics. > (3) infinity is a useful fiction, but emphatically, it > is a fiction. When we create this fiction, we must be > careful that every object living in the world of the > infinite has corresponding observable approximations. Infinity means not finite. If you refuse to accept that there is anything that is not finite, I ask you: how many integers are there? An upper bound will be sufficient, but it must be finite. > And, to repeat myself because so many of my detractors > never seem to catch on, formal systems themselves are > objects that live in the world of mathematics as I > envision it. What makes you think they aren't already? > Those who apply mathematics would have no trouble > living with these requirements. Unless those requirements destroyed the foundations of mathematics and made it impossible to do anything with computations knowing that they worked as expected. This is why I suggest you start doing the work, rather than telling us about what you would like to see done. Be more than a visionary, help build your vision. -- Will Twentyman email: wtwentyman at copper dot net Subject: Re: Die Cantor Die > The basics ideas I stress are: (1) mathematics can be > treated as a science which requires statements to > have falsifiable implications > Falsifiable against what? The world of computation. Let me give a simple example. Goldbach's Conjecture makes a prediction that every even number >2 is the sum of two primes. That is falsifiable in the sense that if we were to pick some even number, and verify by computation that there is no pair of primes (each less than our number) whose sum is equal to our number, we would have falsified the conjecture. > The best you can do is say that a certain > section of mathematics is not a model of reality. You're not following my arguments. Maybe it's my fault. > This is why I suggest you start doing the work, > rather than telling us about what you would like to see done. Be more > than a visionary, help build your vision. Good advice is always welcome. Thanks. Subject: Re: Die Cantor Die >>>The basics ideas I stress are: (1) mathematics can be >>>treated as a science which requires statements to >>>have falsifiable implications >>Falsifiable against what? > The world of computation. Isn't mathematics the mechanism of computation? I am concerned that you may enter a circular loop here. Mathematics can be treated as a science which requires statements to have falsifiable implications against the world of computation which is done in mathematics. What happens if the computation you are testing a new idea against is an incorrect computation? What if one of your estimates is off? > Let me give a simple example. Goldbach's Conjecture makes > a prediction that every even number >2 is the sum of two > primes. That is falsifiable in the sense that if we were to > pick some even number, and verify by computation that there > is no pair of primes (each less than our number) whose sum > is equal to our number, we would have falsified the conjecture. And if you find yourself in the state we are in now, would you estimate that it is true, or leave it as an open problem? Also, how does this different from the current state of affairs? >>The best you can do is say that a certain >>section of mathematics is not a model of reality. > You're not following my arguments. Maybe it's my fault. Then please clarify. -- Will Twentyman email: wtwentyman at copper dot net Subject: Re: Die Cantor Die >> I view the purpose of mathematics as extending the bounds of logical >>consequences. Some of these results have useful application in modeling >>aspects of the real world, but that is not necessarily the intended >>function of any given branch of mathematics. > As I have pointed out, formal systems themselves are objects that > live in the world of computation. What's being questioned is the > interpretation of the formal sentences which are provable within > those formal systems. That is, mathematical statements can have > meaning, but not all statements in all formalisms have meaning. What is the difference between meaning and meaning? > In particular, the assertion that there exists a super-infinite > world is being questioned. What does exists mean in that formal > implication of set theory? Since you put exists in quotes, I don't know that you mean what most of us mean. I interpret it to mean a logical constract that may or may not have anything to do with reality. A concept exists. >> To abandon the standard way of thinking about these sets and adopt >>your proposed approach appears to lead to disagreements between the two >>theories, yet you wish to claim they will be basically the same in their >>results. > I am claiming that those who apply mathematics will see little > or no change. Then why change? >>I am unconvinced that your perspective will allow limits in calculus to >>be formally defined, given the problems above. > To do calculus, we would first define the concepts we want to > deal with, with the recognition of the notions of imprecision > and uncertainty, and then abstract out those notions (i.e. consider > the limit as the uncertainty and imprecision go to zero) I am not convinced that you can effectively define limits without a notion of the infinite. In particular, how will you guarantee that certain numbers exist, if you deal only with finite sets? >>I don't know if these are issues you have thought about, but until I can >>see a resolution between standard mathematics and your proposed new >>foundation, it seems unlikely that I will be joining you in advocating a >>world of the finite. > Well, advocating a world of the finite is a distortion of > what I am doing. > I don't want to push you. Maybe someday it will occur to you > that there is a better way than what you are being taught. There may be better ways than what I have studied, but the ways I have studied work as a whole. I am not convinced that your proposed changes will not cause all of mathematics to collapse or at least require enormous work to reconstruct. -- Will Twentyman email: wtwentyman at copper dot net Subject: Re: Die Cantor Die > That is, mathematical statements can have > meaning, but not all statements in all formalisms have meaning. > What is the difference between meaning and meaning? By putting the word in quotes, I was suggesting to the reader that he should give extra throught to the meaning of the word. Mathematical statements can be meaningful in the sense that they make predictions about the outcome of computational experiments. But not all statements in all formalisms are meaninful in that sense. Subject: Re: Die Cantor Die >>>That is, mathematical statements can have >>>meaning, but not all statements in all formalisms have meaning. >>What is the difference between meaning and meaning? > By putting the word in quotes, I was suggesting to the reader > that he should give extra throught to the meaning of the word. I see. I interpreted it as meaning you wanted to add some subtlety to the standard definition without stating what it was. > Mathematical statements can be meaningful in the sense that > they make predictions about the outcome of computational > experiments. But not all statements in all formalisms are > meaninful in that sense. Is that the only sense in which they are meaningful? Is that the only sense in which they are meaningful to you? -- Will Twentyman email: wtwentyman at copper dot net Subject: Re: Die Cantor Die >>>mathematics is more than just a game. There is an underlying >>>reality to it, which is computation. >>There is more than computation. > I'm not sure you've thought this through. > Try to find one single mathematical statement which would be > of interest to those who apply mathematics, which could not be > interpreted as a statement about the world of computation. There is a key concept here that you don't seem to understand: there is a difference between mathematics and applying mathematics. You are interested in only those areas of mathematics that can be applied. Some of us are not. Why should we restrict ourselves to your area of interest? be codified to be verified by a computer proof-checker, it could also be argued that all of math is *already* computation and that we have already achieved your vision. Perhaps your understanding of computation is more limited than ours? Since you are arguing for a change, it appears you do not accept the second option. >>Why should we not be interested in a >>larger view? > The world of fiction is not reasonably called a larger view > than the world of reality. Switching to literature: the world of fiction has driven many of the inventions that have become part of reality. The notion of travel in space was around long before flight was believed to be possible for man. The internet was in its infancy when virtual reality systems were being envisioned in stories. Linking computers to humans was fantasy for a long time, yet is becoming real now. Fiction can give vision to what reality will become. In a similar way pure math can lead to breakthroughs in applied math. The issue is: we don't know in advance if a non-computational aspect of math will lead to an advance in computational math. Formal logic and set theory were developed to verify that other branches of mathematics were valid. What good is a computation if you don't know whether the results are correct or useful? -- Will Twentyman email: wtwentyman at copper dot net Subject: Re: Die Cantor Die > There is a key concept here that you don't seem to understand: there is > a difference between mathematics and applying mathematics. When trying to distinguish between productive lines of inquiry, and mental masturbation, it's a good idea to consider the possible applications of your inquiry. > we don't know in advance if a non-computational aspect of math will lead > to an advance in computational math. There is some truth to this, although mostly you are arguing with your own misinterpretation of what I'm saying. As I have pointed out way too many times, the study of formal systems does fall within the scope of mathematics as the study of phenomena observable in the world of computation. Subject: Re: Die Cantor Die >>There is a key concept here that you don't seem to understand: there is >>a difference between mathematics and applying mathematics. > When trying to distinguish between productive lines of inquiry, > and mental masturbation, it's a good idea to consider the possible > applications of your inquiry. What applications have you generated so far? >>we don't know in advance if a non-computational aspect of math will lead >>to an advance in computational math. > There is some truth to this, although mostly you are arguing with > your own misinterpretation of what I'm saying. As I have pointed > out way too many times, the study of formal systems does fall within > the scope of mathematics as the study of phenomena observable in > the world of computation. In which case, all current mathematics already falls under your proposal and nothing need change for you to be satisfied. What is it you don't like again? Hint: the best you've done is state a few axioms you disagree with, but you haven't shown how you would formalize replacements (if any), nor have you shown how you would patch together the holes that would be left. As others have observed, your system will preserve arithmatic, but will lose the ability to prove certain other useful statements relating to computation. While I don't know which statements those are, it seems like it would be worthwhile for you to research that, analyze the value of what would be lost, and perhaps start formalizing your ideas. So far I've seen a great deal of rhetoric of a category similar to that produced by politicians. Mathematics, regardless of how you do it, must be a little bit clearer. Right now you have failed to provide a replacement for set theory. Until you can at least offer the axioms you wish to use, it is difficult to take you seriously, especially if this is the progress you have made in the past 15 years. -- Will Twentyman email: wtwentyman at copper dot net Subject: Re: Die Cantor Die > My guess is that you object to the Axiom of Infinity, which basicly says > there is a set containing the natural numbers, and would prefer it to be > that the natural numbers form a proper class. If so, then go ahead and > develop a new set theory with an Axiom of Finitude or similar. Or it > may already exist. Actually, what is wrong with doing this? Other than removing a large part of intresting mathematics. Does it actually damage our ability to deal with numbers in some way? Subject: Re: Die Cantor Die |> My guess is that you object to the Axiom of Infinity, which basicly says |> there is a set containing the natural numbers, and would prefer it to be |> that the natural numbers form a proper class. If so, then go ahead and |> develop a new set theory with an Axiom of Finitude or similar. Or it |> may already exist. | |Actually, what is wrong with doing this? Other than removing a large |part of intresting mathematics. Does it actually damage our ability to |deal with numbers in some way? If in Zermelo-Frankel set theory (as well as various others, I think) you replace the axiom of infinity with an axiom saying that every set is finite the resulting system is basically equivalent to the axioms of elementary arithmetic. There's a sort of obvious model, and the elements can be encoded by integers. By Goedel's theorem there are statements of arithmetic that can't be proven in this system. Some of them can be proven in ZF, however. Keith Ramsay Subject: Re: Die Cantor Die Discussion, linux) > If in Zermelo-Frankel set theory (as well as various others, I think) you > replace the axiom of infinity with an axiom saying that every set is finite > the resulting system is basically equivalent to the axioms of elementary > arithmetic. There's a sort of obvious model, and the elements can be > encoded by integers. > By Goedel's theorem there are statements of arithmetic that can't be > proven in this system. Some of them can be proven in ZF, however. Any known interesting ones? -- Jesse Hughes Surround sound is going to be increasingly important in future offices. -- Microsoft marketing manager displays his keen insight Subject: Re: Die Cantor Die > Any known interesting ones? This is a recurring question. It all depends on what one finds interesting. There are no ordinary-style number-theoretic theorems of ZF (Dirichlet's theorem, Fermat, and so on) which are known to be unprovable in ZF without infinity (or, equivalently, in PA). There are however (apart from metamathematical statements) many combinatorial principles in finite mathematics which are known not to be provable in ZF without infinity, beginning with the Ramsey principle used in the Paris-Harrington theorem. Subject: Re: Die Cantor Die >> My guess is that you object to the Axiom of Infinity, which basicly >> says there is a set containing the natural numbers, and would prefer >> it to be that the natural numbers form a proper class. If so, then go >> ahead and develop a new set theory with an Axiom of Finitude or >> similar. Or it may already exist. > Actually, what is wrong with doing this? Other than removing a large > part of intresting mathematics. Does it actually damage our ability to > deal with numbers in some way? Nothing is wrong with doing this. It would be more useful than Petry's constant discussions about the problems of set theory as it is usually taught. What I propose, ultimately, is that he do some work rather than berate mathematicians for not accepting his ideas. He may find that what he thinks will happen doesn't. I don't particularly care what set of axioms a person chooses to investigate. I do care when they try to suggest, without doing the investigation, that one is somehow superior to another. -- Will Twentyman email: wtwentyman at copper dot net Subject: Re: Die Cantor Die > I'm not sure you've thought this through. > Try to find one single mathematical statement which would be > of interest to those who apply mathematics, which could not be > interpreted as a statement about the world of computation. Elliptic geometry is consistent. It's of great intrest to physicists. Of course your use of the word interpreted gives you pretty much free reign, as any thing can be interpreted as anything else given enough effort. Subject: Proof - erasure trapping technique Can someone please help me to proof the following: Theorem/Proposition: When an erasure polynomial e(x) = a_ix^i + a_{h-1}x^{h-1} + a_{h-2}x^{h-2} + ... + a_0 is added to a codeword c(x) = b_{n-1}x^{n-1} + b_{n-2}x^{n-2} + ... + b_0 in C(x) (C(x) is an (n,k) MDS code generated by g(x)) where a_l = b_l or a_l = 0, e(x) can be determined by adding r(x) to s * c_1(x), where c_1(x) = x^i + d_{h-1}x^{h-1} + d_{h-2}x^{h-2} + ... + d_0 in C(x) and s is a scalar. Here, r(x) = f_{h-1}x^{h-1} + f_{h-1}x^{h-2} + ... + f_0 is the coset representative (or the remainder polynomial) of the received polynomial v(x) = c(x) + e(x), i.e. r(x) = v(x) mod g(x). Furthermore, c_1(x) can be computed as x^i + x^i mod g(x). The scalar s can be found as d_j/f_j where the corresponding b_j is known. We assume that the number of erasures does not exceed d_{min} - 1. (h = n-k = d_{min - 1}) (Thus, the erasure polynomial consists of a single burst erasure of length n-k, and one other single erasure. The received polynomial v(x) has zeros where the erasures occurred. To get the polynomial c1(x), we encode x^i, where x^i is the highest erased coefficient.) Proof: One way to proof this that I thought of was to show that a polynomial p(x) = s*c1(x) + r(x) has the following properties: 1.) p(x) belongs to the same coset as r(x) => H*p^t = H*c1^t + H*r^t = 0 + r^t Thus, p(x) belongs to the same coset as r(x). Here, H is a parity-check matrix of C(x) 2.) p(x) has coefficients at the corresponding places where the erasures occurred: This is a bit tricky for me and I don't know how to accomplish it. I thought of using the parity check matrix again, where H = [h_0,h_1, ..., h_{n-1}], h_i the i-th column of H. Any help and/or suggestions will be greatly appreciated Jaco Versfeld Subject: Re: Hyperelliptic curves examples!! > I would like to some examples of HyperElliptic Curves (HEC) and > supersingular curves too.. of different genus along wiht the points > for different fields.. so I can calculate the divisors and test the > arithmetic algorithm that I have written. Can some one give me > references as to where I can find them. > I thought the term supersingular only applied to elliptic curves. > (so of genus 1). > What does it mean to say that a general curve is supersingular? Supersingular applies to abelian varieties of any dimension, so possibly the OP wants curves whose Jacobian varieties are supersingular. But I suspect that the OP meant to say superelliptic, which is Serge Lang's term for curves of the form y^n=f(x). In more prosaic terms, these are cyclic covers of the projective line. JS Subject: Re: Cardano's formulas >> What's the practical use of Cardano's formula?I don't see any. >> Then,why do we have to learn them? >Just a week or two ago someone in this group was complaining that he >DIDN'T learn Cardano's formulas in school, but they should have >been included. High school math teaching is really moving fast. -- I'm not interested in mathematics that might have anything to do with reality. -- Easterly, in sci.math Subject: Bell curve distribution I'd like to ask a favor of you guys. I need an algorithm that will analyze any set of data and put in into a normal bell-curve distribution, by frequency groups. Or at the very least, I need tips to figure out such an algorithm (I'm not entirely mathematically ignorant, but I'm not exactly highly educated either). In case what I'm looking for is not quite clear, let me explain. I need to be able to place a set of data into a bell curve and then break the bell curve up into a set number of frequency groups, so that I have the value range for each group. Here's a sample set: Group 1 -- data points: 9; value range: 133-183 Group 2 -- data points: 23; value range: 184-388 Group 3 -- data points: 36; value range: 389-603 Group 4 -- data points: 45; value range: 604-822 Group 5 -- data points: 50; value range: 823-1,105 Group 6 -- data points: 45; value range: 1,106-1,590 Group 7 -- data points: 36; value range: 1,591-2,052 Group 8 -- data points: 23; value range: 2,053-2,818 Group 9 -- data points: 8; value range: 2,819-3,807 I don't know how good that sample set is -- that's just an example of what my boss is looking for (that's the sample set he gave me). But, in essence, what I'm looking for is an algorithm that will divide any data set into a distribution similar to that example, and it needs to fall into a normal bell curve as accurately as possible. Any help you guys can give me? -- Mekkala, Atheist #2148 Atheism is ... the bed-rock of sanity in a world of madness. --Emmett F. Fields Subject: Re: mathematical trivia (was: How Did Euler Actually Go About Solving the Bridges of Konigsberg Problem? > Where is this K.9anigsberg? In a small piece of Russia on the Baltic Sea, surrounded by Poland and Lithuania. > What is it called now? Kalingrad. > What country is it now in? Modern-day Russia. > What country was it in during Euler's time? Prussia, I think. Best punishment for bad student: Other students make bad nickname for > student and repeat name in front of class laughing. Good for ego, too. The best punishment for students is to be ignored by their own teachers; as far as I am concerned, when I am a teacher, I'll adopt the following behaviour : 1) The students want to work, I'll do my best to help them. 2) The students don't want to work: I don't care. As far as corporal punishments are concerned (in advanced countries) i guess that if a professor tries to hit a student with a stick, then the student will kick him to death, and i think that's a good thing. In any constitution of any civilized country, there's one thing called the right to keep one's physical integrity intact. (our ancestrors fought to win that right) Subject: Re: a little question of non-math charset=iso-8859-7 Julien Santini > Best punishment for bad student: Other students make bad nickname for > student and repeat name in front of class laughing. Good for ego, too. > The best punishment for students is to be ignored by their own teachers; as > far as I am concerned, when I am a teacher, I'll adopt the following > behaviour : > 1) The students want to work, I'll do my best to help them. > 2) The students don't want to work: I don't care. I think you missed my point entirely. What was meant by make bad nickname, means that society has its very own unique way to punish aberrant school behavior, by using pre-existing means. A student cannot be respected more than what he actually represents in grade school or high school. As school in general is a jungle which will host all sorts of little devils, often those very devils will devise ways to crush other students' egos. I recall from my days in high school that we had a fat student who we called Swan Luncheon Meat. A poor student with 25 degrees of nearsightedness and beer-bottle glasses who needed to stick his face against the blackboard so he could read. This student we called The Telescope. A student who was continuously dosing off, was called Hypnos (Sleep). I was called The Old Man. Everyone got nicknamed according to his social characteristics. My point being that society and nature are relentless. If a kid can survive the kind of psychological abuse during school, it can survive anything. Any other punishment is pretty much unecessary, ideally. > As far as corporal punishments are concerned (in advanced countries) i > guess that if a professor tries to hit a student with a stick, then the > student will kick him to death, and i think that's a good thing. In any > constitution of any civilized country, there's one thing called the right to > keep one's physical integrity intact. (our ancestrors fought to win that > right) Baloney. When I was a grad student in Chicago, kids at the local school were carrying guns. When students are noisy to the point of disrupting the ENTIRE class and not let other students (who wish to) learn, I have a God given right to kick them out of class pronto. There are students in every class who DESIRE to learn. When these students' desire (and right) is inhibited by assholes, I can acquire the right to expel them or use some other justifiable means to keep the class going. That's part of the teacher's job. -- Ioannis Galidakis http://users.forthnet.gr/ath/jgal/ ------------------------------------------ Eventually, _everything_ is understandable Subject: Re: New Knowledge > How, when and why does *New Knowledge* come to humankind? > The 'free-Mickey' lawyer (is his name Lessig?) said that all creativity is > derivative. Darren. > ......ahahaha......but derivate is retro active, Standing on the shoulders of giants is retro active? >hence does not point > to anything that is really novel. No surprise that this notion comes from > a lawyer whose name is Essig = vinegar and who defends copyrights > to last an average of 95 years he was trying to free Mickey, not keep him behind bars. >So, much for an impetus to foster > creativity, looking and searching for the NEW..........ahahahaha.... > OTOH green s would love to have their permit charges and > user fees last for 95 years. But it ain't gonna happen. The creativity > by others that the green turds feed off will be --- OUT SOURCED ---. > ** Outsourcing the only real legacy of environmentalism *** > hanson Subject: Re: Which math tutoring software? > I'm a current college student in Trigonometry. I will be taking > Calculus I, II, III, Differential & Orthogonal Equations. I am also a > soon to be electrical engineering student which is going to require I > know how to solve many different types of problems. What math > software does this group recommend that would help me on problems I > don't know how to solve, by showing me the complete steps to solve the > given problem? If they had software like that, they wouldn't need electrical engineers, would they? They'd just take the pimple farmer out from behind the microphone at Jack In the Box, give him the software, hand him a huge paycheck and whenever he had a problem, he'd just type it in and read the answer back to his boss.....or maybe the boss would just type in the question himself. Subject: Design Question I need to extract and amplify only the components of two signals that are exactly in phase. Out of phase signals of any amplitude are noise. I am thinking this is sort of the opposite of common mode rejection. Should I be pursuing an electronic (hardware) solution or a math (DSP) solution? Is there some way to utilize the phase part of an FFT? Ron H. Subject: Re: Design Question Noise may exist or not exist. You also need an oscilloscope to engineer this. Keep the signals separate and simply amplify. If you want to compare the signals, then the fact they are in phase won't produce cancellation. Common mode rejection occurs because the signals are identical, basically. Subject: Re: Design Question > I need to extract and amplify only the components of two signals that are > exactly in phase. Out of phase signals of any amplitude are noise. I am > thinking this is sort of the opposite of common mode rejection. Should I be > pursuing an electronic (hardware) solution or a math (DSP) solution? Is > there some way to utilize the phase part of an FFT? Phase Locked Loop (PLL)? > Ron H. Subject: Re: Design Question >I need to extract and amplify only the components of two signals that are >exactly in phase. Out of phase signals of any amplitude are noise. I am >thinking this is sort of the opposite of common mode rejection. Should I be >pursuing an electronic (hardware) solution or a math (DSP) solution? Is >there some way to utilize the phase part of an FFT? >Ron H. Isn't that what a lock-in amplifier does? It can pick out a signal that's swamped by orders of magnitude more noise. -- We don't grow up hearing stories around the camp fire anymore about cultural figures. Instead we get them from books, TV or movies, so the characters that today provide us a common language are corporate creatures -- Rebecca Tushnet Subject: Re: Design Question > I need to extract and amplify only the components of two signals that are > exactly in phase. Out of phase signals of any amplitude are noise. I am > thinking this is sort of the opposite of common mode rejection. Should I be > pursuing an electronic (hardware) solution or a math (DSP) solution? Is > there some way to utilize the phase part of an FFT? > Ron H. Google on lock in amplifier (without the quotes). Regards Ian Subject: Re: Design Question >lock in amplifier (without the quotes) >Ian Buckner Owlett's right. Use quote marks for phrases. http://www.google.com/help/refinesearch.html#phrase Subject: Re: Design Question >>lock in amplifier (without the quotes) >>Ian Buckner >Owlett's right. Use quote marks for phrases. >http://www.google.com/help/refinesearch.html#phrase Or used the Advanced Search Page and fill-in the blanks: http://www.google.com/advanced_search?hl=en ...Jim Thompson -- | James E.Thompson, P.E. | mens | | Analog Innovations, Inc. | et | | Analog/Mixed-Signal ASIC's and Discrete Systems | manus | | Phoenix, Arizona Voice:(480)460-2350 | | | E-mail Address at Website Fax:(480)460-2142 | Brass Rat | | http://www.analog-innovations.com | 1962 | Will you still need me, will you still feed me, when I'm sixty-four? Subject: Re: Design Question >>I need to extract and amplify only the components of two signals > that are >>exactly in phase. Out of phase signals of any amplitude are noise. I > am >>thinking this is sort of the opposite of common mode rejection. > Should I be >>pursuing an electronic (hardware) solution or a math (DSP) solution? > Is >>there some way to utilize the phase part of an FFT? >>Ron H. > Google on lock in amplifier (without the quotes). > Regards > Ian lock in amplifier or lock-in amplifier with quotes is even better. Jerry -- Engineering is the art of making what you want from things you can get. [OS lash] [OSl ash] [OSl ash] [OSl ash] Subject: Re: Design Question >>I need to extract and amplify only the components of two signals > that are >>exactly in phase. Out of phase signals of any amplitude are noise. I > am >>thinking this is sort of the opposite of common mode rejection. > Should I be >>pursuing an electronic (hardware) solution or a math (DSP) solution? > Is >>there some way to utilize the phase part of an FFT? >>Ron H. > Google on lock in amplifier (without the quotes). > Regards > Ian Actually I think the result with quotes is better. Three of the first four hits are just what you're looking for. IMHO Subject: Re: Design Question > I need to extract and amplify only the components of two signals that are > exactly in phase. Out of phase signals of any amplitude are noise. I am > thinking this is sort of the opposite of common mode rejection. Should I be > pursuing an electronic (hardware) solution or a math (DSP) solution? Is > there some way to utilize the phase part of an FFT? If E(total) = E(common mode) + E(non-common mode) and a common mode circuit yields E(non-common mode) then E(common mode) = E(total) - E(non-common mode) so you can run the signal through a common mode rejection circuit ( Do a Google search on quadrature rejection ), subtract the resultant signal from the original signal, and then amplify the resultant. Of course, the best way to separate signal from noise, is to cross-correlate the effect signal with the cause signal. A Google search on MTI RADAR might show you how auto-correlation is done with analog circuits. Correlation involves summing the product of two strings of data. If you multiply two random strings of data, and sum the products, you end up with zero in the case of pure noise. The larger the sum of the products, the greater the correlation between the two strings. Note that the time delay (Phase in the case of periodic signals) has to be adjusted to make the data strings overlap, as there is a time interval or phase difference between a cause and an effect. -- Tom Potter http://tompotter.us Subject: Re: Design Question >I need to extract and amplify only the components of two signals that are >exactly in phase. Out of phase signals of any amplitude are noise. I am >thinking this is sort of the opposite of common mode rejection. Should I be >pursuing an electronic (hardware) solution or a math (DSP) solution? Is >there some way to utilize the phase part of an FFT? Take a look at quadrature demodulation - in either hardware or software. It is perfectly suited to extracting the in-phase component of a signal as well as the 90-degrees out-of-phase component. -Robert Scott Ypsilanti, Michigan (Reply through this forum, not by direct e-mail to me, as automatic reply address is fake.) Subject: Re: Design Question > I need to extract and amplify only the components of two signals that are > exactly in phase. Out of phase signals of any amplitude are noise. I am > thinking this is sort of the opposite of common mode rejection. Should I be > pursuing an electronic (hardware) solution or a math (DSP) solution? Is > there some way to utilize the phase part of an FFT? Cross-Correlation Subject: Re: This Week's Finds in Mathematical Physics (Week 202) Here's a neat 18th century way of getting the Catalan numbers. Consider the equation for Lisp S-expressions S = A + S x S meaning that an S-expression is either an atom or a pair of S-expressions. Regard this as a quadratic equation for S and solve it by the quadratic formula. Expand the square root in power series. The coefficients of the powers of A are just the Catalan numbers giving the number of types of S-expression with a given number of atoms. Of course, you get set isomorphisms as well. I suppose this is a special case of John Baez's theory. -- John McCarthy, Computer Science Department, Stanford, CA 94305 http://www-formal.stanford.edu/jmc/progress/ He who refuses to do arithmetic is doomed to talk nonsense. Subject: Re: This Week's Finds in Mathematical Physics (Week 202) X-MailScanner-SpamxCheck: not spam, SpamAssassin (score=-11.3, required 5, Lisp S-expressions > S = A + S x S Lisp has a very simple typing system so there isn't much beyond S=A+SxS. In a language with a type system such as Haskell, where types are built algebraically, then the method used to define types is isomorphic to writing things like S=A+SxS apart from a bit of syntactic sugar. For example a simple tree type in Haskell is defined by data Tree a = Leaf a | Branch (Tree a) (Tree a) The type system in Haskell has been well studied from the point of view of category theory (Curry-Howard isomorphism connecting types with intuitionistic logic etc.) Haskell is also a great computer language for considering some of the other stuff JB has been talking about. For example you can think about a monad that allows you to 'lift' functions on classical systems to functions on quantum systems and then directly implement a quantum computer simulator in Haskell using this monad. http:// www.cs.indiana.edu/~sabry/papers/quantum.pdf (Monads form a fundamental part of Haskell BTW) -- Torque Subject: Re: Role of (Pressure/Volume) quotient in ideal Gas Law [..] > However, I do not find this one variable U=P/V mentioned in physics or > thermodynamics in undergraduate classes. [..] > Can it be agreed that U= P / V is a valid physical quantity ? It looks like you are rewriting some well known laws by rearranging the variables in new combinations. Sometimes this can be useful, depending on what you want to do. Thermodynamics has quite al lot of this, for example, you can express a problem in terms of internal energy, enthalpy, Helmholtz free energy, or Gibbs free enthalpy, depending again on what you want to do. Recently quantities like Exergy have become popular. But I don't see any use for the quantity U (which is usually reserved for internal energy in thermodynamics). You need to show how it can make solving certain problems more easy. Gerard Subject: Re: Differential equation Dear Lee, thank you for this nice posting! > Then we consider the real subspace V of C^2 defined by z_2=(z_1)*, > and call *it* R^2, so that now (in some even more abusive sense) delta(z)delta(z*) means the point mass at the origin of V. Why is this statement abusive? Such kind of hints don't help me, you have to explain to me what actually is wrong with this abusive physicist's notation! Otherwise you maybe satisfied, but I didn't learn anything new. And what is a point mass in the mathematical meaning? Sounds like a term from physics. > Now we consider as our space of test functions something like > the (complex valued) real-analytic functions on V (which of course > are very far from being like nice compactly supported test functions > which you would use to define distributions), any one of which > is of course the restriction to V of a genuine complex-analytic > function on some neighborhood of V in C^2 (and conversely). Why does one need compactly supported test functions to define distributions? In my undergraduate courses we defined the delta distribution as a limit, of for example a properly normalized Gaussian, or a rectangle with Area one. > Even if any of this is right (or rightly interpreted), I too don't > see why you'd want to do it. But then, I don't pretend to understand > physic(ist)s. Well, sometimes, doing some calculations in the coordinates z = x+iy and z* = x-iy is nicer to do than in other coordinates. Ren.8e. -- Ren.8e Meyer Student of Physics & Mathematics Zhejiang University, Hangzhou, China Subject: Re: Differential equation >> Then we consider the real subspace V of C^2 defined by z_2=(z_1)*, >> and call *it* R^2, so that now (in some even more abusive sense) >delta(z)delta(z*) means the point mass at the origin of V. >Why is this statement abusive? The difficulties involved in multiplying distributions have already been noted repeatedly. On top of that, here there's the sleight-of-hand involved in calling V by the name R^2, since R^2 is already present inside C^2 and not equal to V. >Such kind of hints don't help me, >you have to explain to me what actually is wrong with this abusive >physicist's notation! Otherwise you maybe satisfied, but I didn't >learn anything new. And what is a point mass in the mathematical meaning? >Sounds like a term from physics. Well, what *I* meant by it was a measure supported on a singleton, such that the measure of the singleton is 1. Such a measure (or any measure) can be interpreted as a functional on various spaces of functions (and, generally, not on other spaces of functions, or things-like-functions; for instance, you'd have trouble defining delta on L^p for any p, since elements of L^p don't have values at particular points). As soon as one starts slinging terms like distribution around, one is, at least implicitly, making such an identification (it seems to me); but a measure, pure and simple, is not such a functional--at *best*, it's manifested as many different functionals on many different spaces. >> Now we consider as our space of test functions something like >> the (complex valued) real-analytic functions on V (which of course >> are very far from being like nice compactly supported test functions >> which you would use to define distributions), any one of which >> is of course the restriction to V of a genuine complex-analytic >> function on some neighborhood of V in C^2 (and conversely). >Why does one need compactly supported test functions to define >distributions? I dunno. Maybe you don't. It's a matter of terminology, of course. >In my undergraduate courses we defined the delta >distribution as a limit, of for example a properly normalized >Gaussian, or a rectangle with Area one. That would be a limit in what topological (vector)space? Given a measure m which is a point mass at the origin (as I used the term above), and given an appropriate space F of functions (on whatever space the origin is the origin *of*), the functional on F which assigns to a function f in F the the integral of f with respect to m is linear; when F is equipped with a topology, this linear functional may (or may not) be continuous; in that case (or, I suppose, even if it isn't continuous, but is at least defined), you can ask whether this linear functional is (or is not) the limit of an appropriate sequence (or net...) of *other* linear functionals, perhaps defined by integration with respect to *other* measures (but perhaps not). >> Even if any of this is right (or rightly interpreted), I too don't >> see why you'd want to do it. But then, I don't pretend to understand >> physic(ist)s. >Well, sometimes, doing some calculations in the coordinates z = x+iy >and z* = x-iy is nicer to do than in other coordinates. Oh, sure, I know that's true in lots of cases, even cases I have a personal interest in. But I don't know why it's true (for you) in the case being discussed. Lee Rudolph Subject: Re: Differential equation >>> Then we consider the real subspace V of C^2 defined by z_2=(z_1)*, >>> and call *it* R^2, so that now (in some even more abusive sense) >>delta(z)delta(z*) means the point mass at the origin of V. >>Why is this statement abusive? >The difficulties involved in multiplying distributions have >already been noted repeatedly. On top of that, here there's >the sleight-of-hand involved in calling V by the name R^2, >since R^2 is already present inside C^2 and not equal to V. >>Such kind of hints don't help me, >>you have to explain to me what actually is wrong with this abusive >>physicist's notation! Otherwise you maybe satisfied, but I didn't >>learn anything new. And what is a point mass in the mathematical meaning? >>Sounds like a term from physics. >Well, what *I* meant by it was a measure supported on a singleton, >such that the measure of the singleton is 1. Such a measure (or >any measure) can be interpreted as a functional on various spaces >of functions (and, generally, not on other spaces of functions, >or things-like-functions; for instance, you'd have trouble defining >delta on L^p for any p, since elements of L^p don't have values at >particular points). >As soon as one starts slinging terms like distribution around, >one is, at least implicitly, making such an identification (it >seems to me); but a measure, pure and simple, is not such a >functional--at *best*, it's manifested as many different functionals >on many different spaces. Actually, modulo various technicalities, a measure is exactly the same thing as a positive linear functional on the space of continuous functions with compact support. >>> Now we consider as our space of test functions something like >>> the (complex valued) real-analytic functions on V (which of course >>> are very far from being like nice compactly supported test functions >>> which you would use to define distributions), any one of which >>> is of course the restriction to V of a genuine complex-analytic >>> function on some neighborhood of V in C^2 (and conversely). >>Why does one need compactly supported test functions to define >>distributions? >I dunno. Maybe you don't. It's a matter of terminology, of course. There are various sorts of distributions. Roughly speaking, distributions per se give the dual of the smooth functions with compact support; since the functions have compact support there's no restriction on the growth of the distribution at infinity (not that it makes sense to put it exactly that way, since the distribution is not a function.) There's a standard class of rapidly decreasing or Schwarz functions, which are required to die very fast at infinity; the dual of that class is the tempered distributions, which roughly speaking don't grow too fast at infinity. Then one can consider smooth functions with no growth condition at all; the dual of _that_ class is the distributions with compact support. >>In my undergraduate courses we defined the delta >>distribution as a limit, of for example a properly normalized >>Gaussian, or a rectangle with Area one. >That would be a limit in what topological (vector)space? >Given a measure m which is a point mass at the origin (as I >used the term above), and given an appropriate space F of >functions (on whatever space the origin is the origin *of*), >the functional on F which assigns to a function f in F the >the integral of f with respect to m is linear; when F is >equipped with a topology, this linear functional may (or >may not) be continuous; in that case (or, I suppose, even if >it isn't continuous, but is at least defined), you can ask >whether this linear functional is (or is not) the limit of >an appropriate sequence (or net...) of *other* linear functionals, >perhaps defined by integration with respect to *other* measures >(but perhaps not). >>> Even if any of this is right (or rightly interpreted), I too don't >>> see why you'd want to do it. But then, I don't pretend to understand >>> physic(ist)s. >>Well, sometimes, doing some calculations in the coordinates z = x+iy >>and z* = x-iy is nicer to do than in other coordinates. >Oh, sure, I know that's true in lots of cases, even cases I have >a personal interest in. But I don't know why it's true (for you) >in the case being discussed. >Lee Rudolph ************************ Subject: Re: Differential equation ... >>As soon as one starts slinging terms like distribution around, >>one is, at least implicitly, making such an identification (it >>seems to me); but a measure, pure and simple, is not such a >>functional--at *best*, it's manifested as many different functionals >>on many different spaces. >Actually, modulo various technicalities, a measure is exactly >the same thing as a positive linear functional on the space >of continuous functions with compact support. As Herman Rubin often reminds us, no topology (and ipso facto, no continuous functions or compact supports) is/are needed to define a measure, and many interesting measures are defined where topologists fear to tread. Now, if you put the right adjective in front of the word measure (actually, I guess, the right person's last name...but it's been a long time, and I forget), I would agree with you without a quibble. Lee Rudolph Subject: Re: Differential equation >... >>>As soon as one starts slinging terms like distribution around, >>>one is, at least implicitly, making such an identification (it >>>seems to me); but a measure, pure and simple, is not such a >>>functional--at *best*, it's manifested as many different functionals >>>on many different spaces. >>Actually, modulo various technicalities, a measure is exactly >>the same thing as a positive linear functional on the space >>of continuous functions with compact support. >As Herman Rubin often reminds us, no topology (and ipso facto, >no continuous functions or compact supports) is/are needed >to define a measure, and many interesting measures are defined >where topologists fear to tread. Now, if you put the right >adjective in front of the word measure (actually, I guess, >the right person's last name...but it's been a long time, and >I forget), I would agree with you without a quibble. Well yes of course a measure does not require a topology; that person's name (could be Mr regular Borel or Mr Radon depending on who you talk to) was one of the various technicalities I was referring to. (In a context like the present, where we're talking about analysis on R^2, people often just say measure in place of regular Borel measure. _Also_ they often say measure in place of complex Borel measure...) >Lee Rudolph ************************ Subject: Re: Differential equation >>>>You are right. Well, I think this is what I actually forgot! The >>>>manipulation >>>>dd*ln(zz*) = d(1/z*) = d*(1/z) >>> Typo there, of course. >>Where is the typo? > You actually meant dd*ln(zz*) = d(1/z*) = d*(1/z), > not dd*ln(zz*) = d(1/z*) + d*(1/z)? That's not a typo. Have a look at my sketch of the calculation I did. I now know, that ln (zz*) = lnz + lnz* is not true in general as of the problem with the ln of complex numbers. Is there any way to make this equation hold? >>> Another possibility for the inconsistency is that dd* is not exactly >>> the Laplacian, it's the Laplacian times 4 or 1/4 or something. >>No, Polchinski definitely defines dd* as d/dz d/dzbar. > What no? That's how I _assumed_ you were defining dd*. > That's not the Laplacian. You are right, I forgot to give you the definition of d and d*. Its d = d/dz = 1/2(d/dx -i d/dy) and d* the complex conjugate of this, with all derivatives being partial ones. Thus dd* is 1/4 the laplacian. Ren.8e. -- Ren.8e Meyer Student of Physics & Mathematics Zhejiang University, Hangzhou, China Subject: Re: Differential equation >>>>>You are right. Well, I think this is what I actually forgot! The >>>>>manipulation >>>>>dd*ln(zz*) = d(1/z*) = d*(1/z) >>>> Typo there, of course. >>>Where is the typo? >> You actually meant dd*ln(zz*) = d(1/z*) = d*(1/z), >> not dd*ln(zz*) = d(1/z*) + d*(1/z)? >That's not a typo. Right - I realized later I wasn't reading carefully enough. >Have a look at my sketch of the calculation I did. >I now know, that ln (zz*) = lnz + lnz* is not true in general as of >the problem with the ln of complex numbers. Is there any way to make >this equation hold? Like I said, I'm not sure. It's certainly ok away from the origin, but around the origin is the interesting part and that's exactyly where it's not clear to me whether this makes sense or not. >>>> Another possibility for the inconsistency is that dd* is not exactly >>>> the Laplacian, it's the Laplacian times 4 or 1/4 or something. >>>No, Polchinski definitely defines dd* as d/dz d/dzbar. >> What no? That's how I _assumed_ you were defining dd*. >> That's not the Laplacian. >You are right, I forgot to give you the definition of d and d*. Its >d = d/dz = 1/2(d/dx -i d/dy) >and d* the complex conjugate of this, with all derivatives being >partial ones. Thus dd* is 1/4 the laplacian. >Ren.8e. ************************ Subject: Re: Differential equation >>Have a look at my sketch of the calculation I did. >>I now know, that ln (zz*) = lnz + lnz* is not true in general as of >>the problem with the ln of complex numbers. Is there any way to make >>this equation hold? > Like I said, I'm not sure. It's certainly ok away from the origin, but > around the origin is the interesting part and that's exactyly > where it's not clear to me whether this makes sense or not. Actually, the application from where this problem arose assumes r != 0, and phi in [0,2pi] for z = r exp(iphi). Ren.8e. -- Ren.8e Meyer Student of Physics & Mathematics Zhejiang University, Hangzhou, China Subject: Re: Differential equation >>>Have a look at my sketch of the calculation I did. >>>I now know, that ln (zz*) = lnz + lnz* is not true in general as of >>>the problem with the ln of complex numbers. Is there any way to make >>>this equation hold? >> Like I said, I'm not sure. It's certainly ok away from the origin, but >> around the origin is the interesting part and that's exactyly >> where it's not clear to me whether this makes sense or not. >Actually, the application from where this problem arose assumes r != >0, and phi in [0,2pi] for z = r exp(iphi). Um, polar coordinates only make sense for r <> 0. If we're wondering about how to show the laplacian of log(zz*) is (whatever constant times) a point mass at the origin then the problem is _at_ the origin. If we were only talking about points away from the origin the problem wouldn't come up in the first place; the laplacian of log(zz*) is obviously 0 away from the origin. >Ren.8e. ************************ Subject: Re: Differential equation >>> How to show this? >>I looked up a text on distributions :-) > One can give a cheap-hack sort of argument for this, > first expressing the laplacian in polar coordinates > (then recalling the _definition_ of the laplacian of > a distribution) and integrating by parts. And what is that definition? I really have not much knowledge about distribution theory, maybe some of you can enlighten me a little bit ;-) Ren.8e. -- Ren.8e Meyer Student of Physics & Mathematics Zhejiang University, Hangzhou, China Subject: Re: Differential equation >>>> How to show this? >>>I looked up a text on distributions :-) >> One can give a cheap-hack sort of argument for this, >> first expressing the laplacian in polar coordinates >> (then recalling the _definition_ of the laplacian of >> a distribution) and integrating by parts. >And what is that definition? I really have not much knowledge about >distribution theory, maybe some of you can enlighten me a little bit Say u is a distribution and L is the Laplacian. By definition Lu is the distribution such that = is more proper notation for what people sometimes write as int(u phi). Another correct alternative is u(phi).) >;-) >Ren.8e. ************************ Subject: Re: Differential equation >>> It's well-known that -(log|z|)/2 pi >>> is a fundamental solution of the Laplacian: i.e. >>> Delta(log |z|) = -2 pi delta --- usual delta distribution! >>> So d d*(log |z|^2) = -pi delta. >> How to show this? > I looked up a text on distributions :-) Can you tell me some good ones? I never learned much about distribution theory except the stuff one learns as a physicist. Ren.8e. -- Ren.8e Meyer Student of Physics & Mathematics Zhejiang University, Hangzhou, China Subject: Re: Differential equation >>>> It's well-known that -(log|z|)/2 pi >>>> is a fundamental solution of the Laplacian: i.e. >>>> Delta(log |z|) = -2 pi delta --- usual delta distribution! >>>> So d d*(log |z|^2) = -pi delta. >>> How to show this? >> I looked up a text on distributions :-) >Can you tell me some good ones? I never learned much about distribution >theory except the stuff one learns as a physicist. You can find a fairly gentle introduction in Folland Real Analysis. >Ren.8e. ************************ Subject: Re: Differential equation >>>> It's well-known that -(log|z|)/2 pi >>>> is a fundamental solution of the Laplacian: i.e. >>>> Delta(log |z|) = -2 pi delta --- usual delta distribution! >>>> So d d*(log |z|^2) = -pi delta. >>> How to show this? >> I looked up a text on distributions :-) > Can you tell me some good ones? The best is Hormander's volume 1. :-) -- Robin Chapman, www.maths.ex.ac.uk/~rjc/rjc.html Lacan, Jacques, 79, 91-92; mistakes his penis for a square root, 88-9 Francis Wheen, _How Mumbo-Jumbo Conquered the World_ Subject: matlab matrix problem I have a two dimensional matrix U which I'd like to power with a. a has different values depending on it's index a(k). The result should be a three dimensional matrix (the two dimensional matrix powered by a) for each of the 100 different values of a(k). The following program causes the error message: ??? In an assignment A(I) = B, the number of elements in B and I must be the same. for k=1:100 H(k)= U.^a(k); end What is my mistake? any help is appreciated Nicolas Subject: Re: matlab matrix problem > I have a two dimensional matrix U which I'd like to power with a. a has > different values depending on it's index a(k). > The result should be a three dimensional matrix (the two dimensional matrix > powered by a) for each of the 100 different values of a(k). > The following program causes the error message: ??? In an assignment A(I) > = B, the number of elements in B and > I must be the same. It's a matrix-table allocation problem. If H doesn't already exist, the first assignment when k=1, makes H a two-dimensional matix, by defauilt. H=U.^a(1). The usual correction is to define H before you use it in a loop. So use one of Matlabs matrix definition statements like: H=zeros(size(U),100); to preallocate the matrix header table. > for k=1:100 > H(k)= U.^a(k); > end > What is my mistake? > any help is appreciated > Nicolas Subject: Re: matlab matrix problem > I have a two dimensional matrix U which I'd like to power with a. a has > different values depending on it's index a(k). > The result should be a three dimensional matrix (the two dimensional matrix > powered by a) for each of the 100 different values of a(k). > The following program causes the error message: ??? In an assignment A(I) > = B, the number of elements in B and > I must be the same. > for k=1:100 > H(k)= U.^a(k); > end > What is my mistake? > any help is appreciated > Nicolas If matrix H is already 3 dimensional, with first 2 dimensions corresponding to U: for k=1:100 H(:,:,k)= U.^a(k); end Jeroen Subject: Re: Candidate looking for Employment > am looking for a job. Here's a copy of my resume > Name : Suresh kumar Devanathan > contact : mdsuresh ___@thing__ media __.thing__ mit __.thing__ edu Really want a job? Lose the ethnicity. Change your name to John Smith, and be sure not to include a photo with your resume. Best to present them with your ethnicity at the same time you're dazzling them with your brilliant personality. That's just a cynical suggestion, but make sure your resume is picture perfect because that's what will get you the interviews. > Experience : MIT Media Lab > Rutgers CAIP Lab Good magic letters MIT, but doing what? > Awards : AP Scholar With Distinction > Other : Army Reserve Do you still have an obligation? Could you be called back up and sent to Iraq next week? Employers won't want to hear this. In that case it's best not to highlight this. > Worked for : Dr. Vishwani Agarwal( IEEE/ACM fellow) on VLSI ATPG > Dr. Edward Fredkin( CMU Distinguished professor) on > Cellular Automata/ SARS-TA Employers might want to know, What the hell is that! > Public Projects: > http://www.sourceforge.net/projects/atpg > ----------------------------------------- > Built this project, as a PROOF of CONCEPT for statistical > techniques, i developed > to test chips > http://www.sourceforge.net/projects/blitz > ----------------------------------------- > Patch work for blitz, a highly successful C++ math library > in use, by more than > 15000 programmers worldwide Go into a bit more detail. > Job Skill : C++, C#, Unix, Cadence Design System, MATLAB, .NET How many years paid experience in each? > Education : 3 years of college in Electrical Engineering > Left college to jumpstart a partly successful > entrepreneurship Partly successful? Doesn't sound good. You may have to develop a strategy for peddling this. They will want to know the real reason you dropped out of school. And be prepared to supply a transcript. Three years of college usually just tells people you didn't finish. Of course McDonalds is always hiring. No questions asked! Double-A Subject: Question about associativity of cartesian product As I understand it (according to http://en.wikipedia.org/wiki/Cartesian_product), 1. the cartesian product of two sets gives a set of 2-tuples (ordered pairs), i.e. X*Y={(x,y):x in X and y in Y} 2. generally, the cartesian product of n sets gives a set of n-tuples 3. cartesian product is associative. But that third statement seems to contradict the first two: by the first two, (A*B)*(C*D) should give a set of pairs of sets of pairs, but by the third statement, that same expression should give a set of 4-tuples. Obviously I'm missing something obvious. Any enlightenment would be appreciated. Subject: Re: Question about associativity of cartesian product Adjunct Assistant Professor at the University of Montana. >As I understand it (according to >http://en.wikipedia.org/wiki/Cartesian_product), >1. the cartesian product of two sets gives a set of 2-tuples (ordered >pairs), i.e. X*Y={(x,y):x in X and y in Y} >2. generally, the cartesian product of n sets gives a set of n-tuples >3. cartesian product is associative. >But that third statement seems to contradict the first two: by the >first two, (A*B)*(C*D) should give a set of pairs of sets of pairs, >but by the third statement, that same expression should give a set of >4-tuples. >Obviously I'm missing something obvious. Any enlightenment would be >appreciated. While (A*B)*(C*D) is not equal to A*B*C*D (the set of 4-tuples), there is a natural identification that can be made by removing inner parenthesis: identify the element ((a,b),(c,d)) with the element (a,b,c,d). Using this identification, we think of things like A*(B*C) as being the same as (A*B)*C, etc. -- ============================================================== ======== It's not denial. I'm just very selective about what I accept as reality. --- Calvin (Calvin and Hobbes) ============================================================== ======== Arturo Magidin magidin@math.berkeley.edu Subject: Re: Question about associativity of cartesian product >As I understand it (according to >http://en.wikipedia.org/wiki/Cartesian_product), >1. the cartesian product of two sets gives a set of 2-tuples (ordered >pairs), i.e. X*Y={(x,y):x in X and y in Y} >2. generally, the cartesian product of n sets gives a set of n-tuples >3. cartesian product is associative. >But that third statement seems to contradict the first two: by the >first two, (A*B)*(C*D) should give a set of pairs of sets of pairs, >but by the third statement, that same expression should give a set of >4-tuples. Correct. >Obviously I'm missing something obvious. Any enlightenment would be >appreciated. They are not equal, they are *isomorphic*, that is there is a *canonical, uniquely defined* isomorphism (that is, a set bijection) between those two cartesian products. Associativity is to be understood in this sense, that is (A*B)*C = A*(B*C) where = means there exists a bijection (uniquely defined) between the LHS and the RHS. With my best regards, G. Rodrigues Subject: Re: Cantor Paradox by support1.mathforum.org (8.11.6/8.11.6/The Math Forum, $Revision: 1.9 primary) id i2CDZcR10763; >> As there are several proofs of The uncountability of the set of reals, >> what is the point of criticising just one proof? > >> The diagonal argument is the best known and it has been >> used to prove a lot of things besides the uncountability of >> the real numbers. It is important to know what it can prove >> and what it can't prove. > > >> >> - 2 many 2 count > > >What theorem, if any, has it been used to prove that has no other proof? >If, as I suspect, none, then what is the point of criticizing it? Virgil, you ask why if a conjecture doesn't actually prove a theory but in spirit implies the same result, why it is not just fine. You question the criticism of such a result when another method implies the same result. The point of criticism of one is that, assuming there are two, that both must be criticized. With the antidiagonal argument in considering mapping the reals and naturals, the reals have dual representation, and the binary case is sufficient and required. Applying the antidiagonal argument to an infinite binary sequence for each subset of the naturals in illustrating the set/powerset result is a different result, where there is no dual representation of a subset as an infinite binary sequence. In the consideration of the nested intervals and the gaplessness of the reals, in an effort to show that inexistence of mapping reals to naturals, there are certain piecewise functions, of a sort, that would not have that result, as was addressed and attributed to iota as a least positive real. Criticism of the antidiagonal result is warranted because people argue it in- correctly, acceptance of dual representation of reals show that any list of reals in the unit interval for simplicity could be reordered to have an anti- diagonal that is actually a list element. The Sun rises each day because the Earth orbits the Sun, not the other way around except nominally as the Sun has mass much larger than the Earth as they spin in the cosmic clockworks. The Sun does rise and set each day, that's the definition of a day. Applying that concept to the powerset as represented by the infinite binary sequence is actually quite a different approach than application to the set of reals. If a) the antidiagonal argument and b) the nested intervals arguments are each found to be flawed, then, that doesn't necessarily imply a flaw in an argument using still another method, but it does mean that they can not be used to sup- port that theory. The other argument of course is that a set can not map bijectively to its own powerset and that any set that maps bijectively to its powerset can not map bijectively to it or else through composition of functions the set would map to its powerset. In a model allowing the consideration of infinitesimal real numbers directly addressable as elements of the range or coimage of the function with the do- main of the naturals, and also one where reals have dual representation as binary sequences, there is not a result but Cantor-Bernstein argument by con- tradiction. In that case, various criticisms of the powerset result itself are able to focus on that result without false claims of other non-results, or rather ably to easily dispatch those claims. Some of the considerations of the set and its powerset go in the direction of a set as ordinals, with multiple meanings of their structure and inter- pretation, and with as well considerations of the recomposition of the sets according to structural and ordering properties of their elements. The Cantor Paradox is that the set of all sets would be its own powerset and thus map bijectively to itself by the identity function. If a set is not otherwise enjoined from mapping its own powerset then that result of the universal set being its own powerset is not a contradiction. As well, when within ZF the negation of Cantor's powerset theorem is shown, or even axiomatized, trivially, it doesn't necessarily cascade into contradiction. Personally I think a function on the naturals with a range defined on [0,1], a monotonically increasing function, is an excellent and even fundamental tool. The problem with that is the expectation that the integration of it would have a value of one half, and explaining why it is that instead of the other value, determined by different methods, actually requires an explanation of the reals as points and the reals as a continuous line. The criticism of the antidiagonal argument as applied to reals as for example through the properties of dual representation of reals and sufficiency and necessity of the binary case in refuting the antidiagonal argument in that way is acceptible, or even acceptable, but mostly acceptible. Combined with description of other mappings from the integers to the reals that are not pro- scribed by other results, an entire concept of mapping the integers to specific sets of each element of intervals of the reals is opened for mathematical dis- course free of pre-Galilean prejudices, and variously browbeating, iconoclasty, useless restriction, moribund lassitude, and profound ignorance. Warm regards, Ross F. Subject: Re: Cantor Paradox [already answered] I owe Easterly an apology. I carelessly misread the ravings of Ross A. Finlayson as being by . Sorry. Should have known better. Subject: Re: Cantor Paradox >> As there are several proofs of The uncountability of the set of reals, >> what is the point of criticising just one proof? > >> The diagonal argument is the best known and it has been >> used to prove a lot of things besides the uncountability of >> the real numbers. It is important to know what it can prove >> and what it can't prove. > > >> >> - 2 many 2 count > > >What theorem, if any, has it been used to prove that has no other proof? >If, as I suspect, none, then what is the point of criticizing it? > Virgil, you ask why if a conjecture doesn't actually prove a theory but in > spirit implies the same result, why it is not just fine. > You question the criticism of such a result when another method implies the > same result. The point of criticism of one is that, assuming there are two, > that both must be criticized. > With the antidiagonal argument in considering mapping the reals and naturals, > the reals have dual representation, and the binary case is sufficient and > required. > Applying the antidiagonal argument to an infinite binary sequence for each > subset of the naturals in illustrating the set/powerset result is a different > result, where there is no dual representation of a subset as an infinite > binary sequence. The failure onf one proof for an alleged theorem does not disprove the theaorem, but the success of even one proof does prove it. The issue is not whether there is a base in which Cantor's diagonal construction fails, but whether there is a base in which Cantor's diagonal construction succeeds, as it does for any integral base larger than three. The fact that it might fail with bases two and three is irrelevant. > In the consideration of the nested intervals and the gaplessness of the > reals, > in an effort to show that inexistence of mapping reals to naturals, there are > certain piecewise functions, of a sort, that would not have that result, as > was addressed and attributed to iota as a least positive real. It is mappings from naturals to reals, not the reverse, that is at issue here. Get your facts straight. And none of your blatherings on the subject invalidated the first proof anywhere but in your own fuzzy mind. > Criticism of the antidiagonal result is warranted because people argue it in- > correctly, But it only takes one to argue it correctly, as Cantor himself did, to establish the result. That others do it wrong is of no consequence. acceptance of dual representation of reals show that any list of > reals in the unit interval for simplicity could be reordered to have an anti- > diagonal that is actually a list element. The Sun rises each day because the > Earth orbits the Sun, not the other way around The sun rises each day, NOT because it rotates about the sun, but because it rotates about itself. It is the seasons and years that result from earth's rotation about the sun. 's knowledge of astronomy is on a par with his knowledge of mathematics. except nominally as the Sun > has > mass much larger than the Earth as they spin in the cosmic clockworks. The > Sun does rise and set each day, that's the definition of a day. > Applying that concept to the powerset as represented by the infinite > binary sequence is actually quite a different approach than > application to the set of reals. Which concept? Rotation about the sun? > If a) the antidiagonal argument and b) the nested intervals arguments > are each found to be flawed, then, that doesn't necessarily imply a > flaw in an argument using still another method, but it does mean that > they can not be used to sup- port that theory. But no one has shown either proof to have the slightest flaw, except for the difficulty that the stupid have in understanding them. > The other argument of course is that a set can not map bijectively to its own > powerset and that any set that maps bijectively to its powerset can not map > bijectively to it or else through composition of functions the set would map > to its powerset. I think's short circuits are showing again. > In a model allowing the consideration of infinitesimal real numbers directly > addressable as elements of the range or coimage of the function with the do- > main of the naturals, and also one where reals have dual representation as > binary sequences, there is not a result but Cantor-Bernstein argument by con- > tradiction. I think needs to reboot his internal operating system, and check for viruses. > In that case, various criticisms of the powerset result itself are > able to focus on that result without false claims of other > non-results, or rather ably to easily dispatch those claims. Definitely needs a reboot or virus check. [The rest deleted as incoherent] Subject: Question on modular arithmetic I have a question regarding modular arithmetic: Is it possible to easily form a polynomial form of a modular arithmetic series (much like arithmetic or geometric series)? It'd make the algorithm I'm working on much easier if it is. Subject: Question about combining loop QG + m-theory If you bump up m-theory to 12 dimensions it breaks down into a quantum mechanics m-theory of the microcosm in 12 dimensions and a relativistic m-theory of the macrocosm in 12 dimensions. In order to combine loop quantum gravity with this 12 dimensional approach to m-theory I speculate I would have to cut the 12 dimensions up into 4 three dimensional slices. For example: 3 dimensions of space (length ,width ,depth), 3 dimensions of time(past, present, and future), 3 dimensions of mass (matter, antimatter, supermatter), 3 dimensions of solidity(atomic nucleus, the whole atom, and molecules). The point being you could measure each 3 dimensional slice on an x,y,z coordinate system. my question is: Is there mathematics that lets you stack or combine x,y,z coordinate systems so you can represent 12 dimensions mathematicly in order to combine loop Quantum gravity with m-theory Subject: pi=2.83... ? Was just wondering... I have implemented a classical Bresenham midpoint circle algorithm (using 8-symetry). It works pretty well. I have carefully counted (discarding possible overdraws) the pixels plotted by the function in order to draw a circle of radius r (in pixels). This value is the perimeter of the discretized circle (in pixels), something close from 2*pi*r. By dividing this by 2*r... I may obtain an approximation of pi, but not. Here is my problem, I find out a converging pi value of about 2.83... I may be missing something *very big* Any clues appreciated, Thanks in advance. Subject: Re: pi=2.83... ? In-reply-to: cbergadler@yahoo.fr (Carlos Bergadler) >I have implemented a classical Bresenham midpoint circle algorithm >(using 8-symetry). It works pretty well. >I have carefully counted (discarding possible overdraws) the pixels >plotted by the function in order to draw a circle of radius r (in >pixels). >This value is the perimeter of the discretized circle (in pixels), >something close from 2*pi*r. By dividing this by 2*r... I may obtain >an approximation of pi, but not. >Here is my problem, I find out a converging pi value of about 2.83... >I may be missing something *very big* When computing arclength, you have to be careful that the pieces you are adding actually measure arclength. A pixel does not measure arclength. For example, the line from (0,0) to (100,100) contains approximately 100 pixels. However, the distance between these points is between 141 and 142 pixel widths. On parts of curves that are more horizontal, counting pixels measures the horizontal extent. On parts of curves that are more vertical, counting pixels measures the vertical extent. Between 0 and 45 degrees, a circle is vertical; between 45 and 90 degrees, it is horizontal. Therefore, the number of pixels from 0 to 45 degrees is going to be radius/sqrt(2), and between 45 and 90 degrees, it is also going to be radius/sqrt(2). The circumference of a circle is 8*radius/sqrt(2) if measured this way. Dividing this by a diameter of 2*radius yields 4/sqrt(2) = 2.828427... for pi. Rob Johnson Was just wondering... > I have implemented a classical Bresenham midpoint circle algorithm > (using 8-symetry). It works pretty well. > I have carefully counted (discarding possible overdraws) the pixels > plotted by the function in order to draw a circle of radius r (in > pixels). > This value is the perimeter of the discretized circle (in pixels), > something close from 2*pi*r. By dividing this by 2*r... I may obtain > an approximation of pi, but not. > Here is my problem, I find out a converging pi value of about 2.83... > I may be missing something *very big* Are you in Indiana by any chance? Bart Subject: Re: pi=2.83... ? >Was just wondering... >I have implemented a classical Bresenham midpoint circle algorithm >(using 8-symetry). It works pretty well. >I have carefully counted (discarding possible overdraws) the pixels >plotted by the function in order to draw a circle of radius r (in >pixels). >This value is the perimeter of the discretized circle (in pixels), >something close from 2*pi*r. By dividing this by 2*r... I may obtain >an approximation of pi, but not. >Here is my problem, I find out a converging pi value of about 2.83... >I may be missing something *very big* I'm not sure I understand. You're calculating the perimeter of the circle by counting the number of pixels that the algorithm colors when it draws the circle? If so there's really no reason that should give the perimeter - pixels are not round, the distance between the centers of two adjacent pixels is either 1 or sqrt(2), etc. If that's not what you're doing please explain. >Any clues appreciated, >Thanks in advance. ************************ Subject: Re: pi=2.83... ? Originator: richard@cogsci.ed.ac.uk (Richard Tobin) >Any clues appreciated, Try using the same method with a 45-degree straight line to get an estimate of sqrt(2). (For extra credit, determine why your approximation to pi is roughly 2*sqrt(2).) -- Richard Subject: Division Rings Over Their Centers Dear All, Lets call an integer number n nice if there exists a division ring which is of dimension n (as a vector space) over its center. Now the question is: Which integer numbers are nice!? Thanks. Subject: Re: Division Rings Over Their Centers > Dear All, > Lets call an integer number n nice if there exists a division ring > which is of dimension n (as a vector space) over its center. > Now the question is: Which integer numbers are nice!? The squares! The buzzword is central simple algebras. See, for instance, Peirce, _Associative Algebras_ or Patrick Morandi's notes at http://sierra.nmsu.edu/morandi/notes/mathematicalnotes.html . -- Robin Chapman, www.maths.ex.ac.uk/~rjc/rjc.html Lacan, Jacques, 79, 91-92; mistakes his penis for a square root, 88-9 Francis Wheen, _How Mumbo-Jumbo Conquered the World_ Subject: Re: Is this travelling salesman problem version known? > If you are required to return to the starting point, > then this is the Hamiltonian cycle problem, > which is knownto be NP-hard. No. The path is allowed to be open (and in the problem I'm working on, it is always open except we have only one city) I din't even see a proof of NP-hardness for the open path version of TSP. Here help is appreciated too. -- Best regards, Alex. PS. To email me, remove loeschedies from the email address given. Subject: Re: matrix analysis: can I say matrix X=Y if I have AXA=AYA? >>1. the dimensions of A,X,Y should be compatible, in this case I think >>they should be all square matrices. > Not at all. Right you are! -- Best regards, Alex. PS. To email me, remove loeschedies from the email address given. Subject: E and E{0} are not homeomorphic Hi all, Can anyone explain me how to prove that, given a normed space E, E and E{0} are not homeomorphic? Of course, I am not interested in the finite-dimensional case. Best regards, Jose Carlos Santos Subject: Re: E and E{0} are not homeomorphic >Can anyone explain me how to prove that, given a normed space E, E and >E{0} are not homeomorphic? Of course, I am not interested in the >finite-dimensional case. It can't be proved, because it isn't always true. For instance, Hilbert space H is homeomorphic (indeed, diffeomorphic) to H{0}, and much, much more is true. Start from West, James E., The diffeomorphic excision of closed local compacta from infinite- dimensional Hilbert manifolds. Compositio Math. 21 1969 271--291 and work forwards and backwards through the literature. (If you have broadband or plenty of time, download Kriegel and Michor's free book at http://www.ams.org/online_bks/surv53/surv53.pdf -- it's 4.68 megabytes, loaded with relevant material.) Lee Rudolph Subject: Re: E and E{0} are not homeomorphic >>Can anyone explain me how to prove that, given a normed space E, E and >>E{0} are not homeomorphic? Of course, I am not interested in the >>finite-dimensional case. > It can't be proved, because it isn't always true. For instance, > Hilbert space H is homeomorphic (indeed, diffeomorphic) to H{0}, > and much, much more is true. Start from West, James E., The > diffeomorphic excision of closed local compacta from infinite- > dimensional Hilbert manifolds. Compositio Math. 21 1969 271--291 > and work forwards and backwards through the literature. (If you > have broadband or plenty of time, download Kriegel and Michor's > free book at http://www.ams.org/online_bks/surv53/surv53.pdf -- > it's 4.68 megabytes, loaded with relevant material.) Wow! Thanks a lot. I really thought that they were never homeomorphic. Best regards, Jose Carlos Santos Subject: Re: E and E{0} are not homeomorphic >>>Can anyone explain me how to prove that, given a normed space E, E and >>>E{0} are not homeomorphic? Of course, I am not interested in the >>>finite-dimensional case. > >> It can't be proved, because it isn't always true. For instance, >> Hilbert space H is homeomorphic (indeed, diffeomorphic) to H{0}, >> and much, much more is true. ... >Wow! Thanks a lot. I really thought that they were never homeomorphic. Proving that they are diffeomorphic is (or used to be...) rather complicated. It's much less difficult to show that they are homotopy equivalent, and even less difficult to show that they are weakly homotopy equivalent (from which, if I remember aright, the strong homotopy equivalence follows easily with a bit more work). First, in any normed space E, E{0} is homotopy equivalent to the unit sphere S of E (with respect to the norm). Now, when we consider H to be the completion of its subvectorspace R^{infty} (of sequences that are eventually 0), then we have S as the completion of S', the intersection of S with R^{infty}; and S' is the limit (in whatever the right technical sense is) of its finite-dimensional subspheres S^k (where a point of S' is in S^k if, as a sequence, it is 0 from the k+1st entry on). But this limit is (weakly) contractible, because S^k contracts to its basepoint within S' (the trace of its contraction can be taken to be the upper hemisphere of S^{k+1}). Some handwaving and you're done. The proof that H is homeomorphic to H{0} is of intermediate difficulty. I heard Bing lecture on it about 35 years ago. The rough idea (which is all I got out of it at the time) is to construct the homeomorphism from H{0} onto H as the limit of a sequence of homeomorphisms f_i from H{0} onto the open subset U_i = H{v_i}, where, in terms of the standard basis e_1, e_2, ..., e_n, ... of H, v_i=e_1+....+e_i. Since the sequence {v_i} doesn't converge in H, in the limit you get a map f from H{0} onto H. Obviously there's something tricky going on there (and you have to be careful to make f exist, and to be a homeomorphism, and to be onto), but when Bing waved his hands it was very convincing to the younger me. Lee Rudolph Subject: (Nonisomorphic )Groups charset=iso-8859-1 According to my book, there are 267 groups of order 64 (we don't demand that they're isomorphic). Which are they? I have found 11 only and i'd be delighted if somebody gave me a hint on how to find the rest. I choose not to go insane and use Z3 x Z5 x Z11 denoted as [3 5 11]. I have found following groups: [2 2 2 2 2 2] [4 2 2 2 2] [8 2 2 2] [16 2 2] [32 2] [64] [4 4 2 2] [4 4 4] [8 4 2] [8 8] [16 4] Even if i go wild and count [16 2 2] and [2 16 2] as different groups i still score 32, not as the book claims 267. I don't need (don't want) a full list, of course but a hint on what groups i missed. -- Kindly Konrad --------------------------------------------------- May all spammers die an agonizing death; have no burial places; their souls be chased by demons in Gehenna from one room to another for all eternity and more. Sleep - thing used by ineffective people as a substitute for coffee Ambition - a poor excuse for not having enough sense to be lazy --------------------------------------------------- Subject: Re: (Nonisomorphic )Groups > According to my book, there are 267 groups of order 64 (we don't demand > that they're isomorphic). Which are they? > I have found 11 only and i'd be delighted if somebody gave me a hint on > how to find the rest. > I choose not to go insane and use Z3 x Z5 x Z11 denoted as [3 5 11]. I > have found following groups: > [2 2 2 2 2 2] [4 2 2 2 2] [8 2 2 2] > [16 2 2] [32 2] [64] [4 4 2 2] > [4 4 4] [8 4 2] [8 8] [16 4] > Even if i go wild and count [16 2 2] and [2 16 2] as different groups i > still score 32, not as the book claims 267. I don't need (don't want) a > full list, of course but a hint on what groups i missed. You simply missed all non-abelian groups - your count for the abelian ones seems to be correct. -- I'm on warm milk and laxatives Cherry-flavored antacids reverse my forename for mail! - saibot Subject: Re: (Nonisomorphic )Groups charset=iso-8859-1 > You simply missed all non-abelian groups - your count for the abelian ones > seems to be correct. I suspected it might be something like that. The problem is that my book is a little bit unclear on that subject (that, or i'm too limited to get that part). For instance - Z(125) has following abelian groups (since 125 = 5*5*5): Z(5^3), Z(5^2) x Z(5) and Z(5) x Z(5) x Z(5). According to the printed thing there are two non-abelian. Nice, i think to my self, but which are they?! I can only think of Z(5) x Z(5^2) but that ought to be incorrect since i'm still one group short... -- Kindly Konrad --------------------------------------------------- May all spammers die an agonizing death; have no burial places; their souls be chased by demons in Gehenna from one room to another for all eternity and more. Sleep - thing used by ineffective people as a substitute for coffee Ambition - a poor excuse for not having enough sense to be lazy --------------------------------------------------- Subject: Re: (Nonisomorphic )Groups >> You simply missed all non-abelian groups - your count for the abelian >> ones seems to be correct. > I suspected it might be something like that. The problem is that my book > is a little bit unclear on that subject (that, or i'm too limited to get > that part). > For instance - Z(125) has following abelian groups (since 125 = 5*5*5): > Z(5^3), Z(5^2) x Z(5) and Z(5) x Z(5) x Z(5). > According to the printed thing there are two non-abelian. Nice, i think to > my self, but which are they?! > I can only think of Z(5) x Z(5^2) but that ought to be incorrect since > i'm still one group short... All groups of the sort Z(n_1) x ... x Z(n_k) are abelian because they are products of abelian groups; to get one of the many non-abelian groups of order 64, take for example the dihedral group D_32. (If you want, there exist arbitrary products in the category of abelian groups which are preserved under the canonical injection functor AbelianGroups ---> AllGroups. BTW, are products, coproducts and limits always conserved under functors? ;-) -- I'm on warm milk and laxatives Cherry-flavored antacids reverse my forename for mail! - saibot Subject: Re: (Nonisomorphic )Groups charset=iso-8859-1 Thanks to both you and Arturo for the help. I see more than clearly where i confused myself. Also, my interest in going deeper into algebra has most certainly increased. Thanks! -- Kindly Konrad --------------------------------------------------- May all spammers die an agonizing death; have no burial places; their souls be chased by demons in Gehenna from one room to another for all eternity and more. Sleep - thing used by ineffective people as a substitute for coffee Ambition - a poor excuse for not having enough sense to be lazy --------------------------------------------------- Subject: Re: (Nonisomorphic )Groups Adjunct Assistant Professor at the University of Montana. >> You simply missed all non-abelian groups - your count for the abelian ones >> seems to be correct. >I suspected it might be something like that. The problem is that my book is >a little bit unclear on that subject (that, or i'm too limited to get that >part). Describing all the groups of order 2^n (or p^n, p a prime) is highly nontrivial. See the review quoted at the end. >For instance - Z(125) has following abelian groups (since 125 = 5*5*5): >Z(5^3), Z(5^2) x Z(5) and Z(5) x Z(5) x Z(5). >According to the printed thing there are two non-abelian. Nice, i think to >my self, but which are they?! >I can only think of Z(5) x Z(5^2) but that ought to be incorrect since >i'm still one group short... Z(5) x Z(5^2) is abelian. You already counted it: it is isomorphic (obviously so) to Z(5^2) x Z(5). If p is a prime, then there are exactly two nonisomorphic nonabelian groups of order p^3. If p = 2, you get the dihedral group of order 8 D_8 = If p is a prime, then there are exactly two nonisomorphic nonabelian > groups of order p^3. > If p = 2, you get the dihedral group of order 8 > D_8 = and the quaternion group of order 8: > Q_8 = <1,-1,i,-i,j,-j,k,-k : i^2 = j^2 = k^2 = -1, (ijk)^2 = 1 I've usually seen this with i^2 = j^2 = k^2 = ijk = -1. Although I may just be carrying over a bit too much from the non-abelian extension of the reals. -- Daniel W. Johnson panoptes@iquest.net http://members.iquest.net/~panoptes/ 039 53 36 N / 086 11 55 W Subject: Re: Cantor Paradox by support1.mathforum.org (8.11.6/8.11.6/The Math Forum, $Revision: 1.9 primary) id i2CFItI24904; >> Wow, sounds like Nathan has rediscovered Richard's Paradox (1905). Not >> bad for an 11-year-old, if he really did think of himself. >He's not an 11-year-old. Whether he's ever read Richard's Paradox or >not, he clearly has the background to choose the issues carefully >here. >(I hadn't seen the paradox before, so I appreciate having a name >attached to it.) >-- > >Contrariwise, continued Tweedledee, if it was so, it might be, and >if it were so, it would be; but as it isn't, it ain't. That's logic! > -- Lewis Carroll I was going to dispute the fact that Nathan's paradox is identical to Richard's Paradox, but now I believe they are identical. I also really appreciate knowing the correct name for this fallacious but very amusing argument. Thanks, Fred Galvin. I used to read the news group geometry_research a lot, and one of the delightful aspects of that forum was that often John Conway would supply the correct names for things. Dr. Conway is also highly adept at coining new names as the occasion demands. Danny Purvis Subject: Re: Cantor Paradox Danny Purvis says... >I was going to dispute the fact that Nathan's paradox is identical to >Richard's Paradox, but now I believe they are identical. On this web page, Nathan's paradox is explicitly connected with Richard's paradox: http://www.dpmms.cam.ac.uk/~wtg10/richardsparadox.html -- Daryl McCullough Ithaca, NY Subject: Re: stochastic group S(2,4) by support1.mathforum.org (8.11.6/8.11.6/The Math Forum, $Revision: 1.9 primary) id i2CFIu424926; >>> How to prove it is isomorphic to alternating group of order 12? >>Um, what is it? >Never heard of it. S(2,4) is sometimes used for symplectic group Sp(2,4), >but that is the same as SL(2,4) and is isomorphic to alternating group of >order 60. >It's not SO^+(2,4) or SO^-(2,4) either - they have orders 6 and 10. >Derek Holt. I've never heard of this group, but could this have something to do with stochastic matrices? These are matrices with all column sums equal to 1, or somehting like that. I don't know how one would make a group form them though. Mark Subject: Traces of matrices I found the following exercise about matrices: show that there are no matrices A and B (over R or even over C) such that AB - BA = I, I the identity matrix. The solution I gave is very simple: For every n x n matrices A and B, we have Tr(AB - BA) = Tr(AB) - Tr(BA) = Tr(AB) - Tr(AB) = 0, where Tr means the trace of the matrix. Since Tr(I) = n>0, there can't be matrices A and B such that AB - BA = I. A similar argument shows that, in fact, if Tr(M)<>0, then there are no matrices A and B such that AB - BA = M. Now, I'm trying to prove, or give a counter example, that if Tr(M) = 0, then there are indeed matrices A and B such that AB - BA = M. I got stuck. I noticed that if D is a diagonal matrix, then, for every matrix A, DA and AD have the same main diagonal, so that the main diagonal of DA -AD is totally formed by 0s. Then I tried to prove that if Tr(M) = 0, then M is similar to a matrix with a null diagonal, but I'm not sure if this is true. Any hints are welcome. Thank yoy Amanda Subject: Re: Traces of matrices > Now, I'm trying to prove, or give a counter example, that if Tr(M) = > 0, then there are indeed matrices A and B such that AB - BA = M. I got > stuck. I noticed that if D is a diagonal matrix, then, for every > matrix A, DA and AD have the same main diagonal, so that the main > diagonal of DA -AD is totally formed by 0s. Then I tried to prove > that if Tr(M) = 0, then M is similar to a matrix with a null diagonal, > but I'm not sure if this is true. Hint: consider the matrices E_{ij} which have only 0's, except for the entry at column i and row j, which is equal to 1. Any M such that Tr(M) = 0 is a linear combination of some E_{ij} with i different from j and some E_{ii} - E_{jj}, again with i different from j. Best regards, Jose Carlos Santos Subject: Re: AC <=> trichotomy > Does anyone have a quick sketch of a proof that AC <=> trichotomy? The => implication follows since AC implies that every set is well ordered (use transfinite induction or Zorn's Lemma) and hence it is easy to find the injection from one set to the other (or just use Zorn's Lemma straight). The <= implication follows, I recall, from something clever called, maybe, Hartogg's (spelling?) Lemma. This says that for any set there exists an ordinal such that there is no injection from the ordinal to the set. Trichotomy then implies that there exists an injection from the set to an ordinal, hence it is well ordered. That every set is well ordered implies AC is very easy. My recollection is that the proof of Hartogg's Lemma is something like this. Given a set A, consider the collection of ordinals which map injectively into the set A. Use the usual set axioms (e.g. replacement and power set) to show that this collection is indeed a set. Hence this collection is itself an ordinal, and since an ordinal cannot be an element of itself, this ordinal satisfies the requirements. Subject: Re: New Randomness Test . > This does not alter the fact that the sequence is not > truly random, since it is entirely predictable given > the knowledge of the appropriate formula. Following this criterion do not exists finite random sequences, because in the development of sqr(2) we can find any given finite sequence of digits. As all we know the program for producing the digits of sqr(2) is very short. L. Rodriguez Subject: Re: New Randomness Test > . > >> This does not alter the fact that the sequence is not truly random, >> since it is entirely predictable given the knowledge of the appropriate >> formula. > > Following this criterion do not exists finite random sequences, That's right. Is a sequence like 1 2 random? It seems to be very predictable but, do we know if it was randomly generated? For finite sequences the best that one can do is to analyze their statistical properties - tapping a truly random source (e.g. a decaying radioactive sample) given enough time will produce any finite sequence, including a long chain of zero bits. The probability for this to happen, however, is minuscule, which is why statiscal tests will raise a red flag: if you happen to encounter such a sequence as produced by a purported source of randomness, chances are that it isn't. Subject: Shafer-Dempster vs. Bayes by support1.mathforum.org (8.11.6/8.11.6/The Math Forum, $Revision: 1.9 primary) id i2CG4dw30832; Dear Friends, I am doing some research about treating uncertainties in engineering calculations. As one method I found the Shafer-Dempster Theory of evidence. I read that resulting from that theory, the Bayesian approach of treating uncertain data is a special case of the Shafer-Dempster Theory and that is the point where I got confused. Can anybody explain me that? I also would like to know what the advantages and disadvantages of both approaches are. Thanks, Dan Subject: Re: Documents about multiplicative order in general and mersenne numbers >> Of course I am not just interested in exponents which are composed >> of exactly two exponents, but in all composed exponents in >> general, because i know now that the divisors of mersenne numbers >> with prime exponents are Cunningham numbers, right? >Sorry, never heard the expression, Cunningham number. > By your address I guess you are a student? Always have been, hope I always will be. > Nobody really knows everything. I would be the last to expect that, > since I didn't even study math, but something else. > If you are interested in Cunningham Numbers you will find documents > here: > http://mathworld.wolfram.com/CunninghamNumber.html So Mathworld defines a Cunningham number as any number of the form b^n + 1 or b^n - 1. But of course that makes every number a Cunningham number, since m = (m + 1)^1 - 1, so help me out a little here. I know what the Cunningham project is, and I know what the Cunningham tables are, but I searched Math Reviews for Cunningham number and came up with nothing. It seems to be mathworld's invention, rather than a standard term. > Mathworld is really a great site. I really can recomment it to you. It is a really great site, but I recommend great caution using it. I've seen too many posts to this newsgroup pointing out errors in mathworld. On the Cunningham number page, it says, Primes of the form b^n - 1 are also very rare. The Mersenne numbers are known to be prime only for 37 values, the first few of which are n = 2, 3, 5, 7, 13, 17, 19, ... (Sloane's A000043). There are no other primes b^n - 1 for nontrivial b < 21 and 1 < n < 1001. You know, I'll bet you can prove that there are no primes b^n - 1 at all for b > 2 and n > 1. So this isn't exactly an error on mathworld, but it's certainly a silly thing for mathworld to say. > In fact I wanted to reply using email, but it didn't work isn't > your correct address? Yes, it is, and I don't know why it didn't work for you, but you know, there's a reason why I only put a disguised form of it in my posts; it has to do with keeping out spam. I'm not happy that you put the undisguised version in your post. > I really got the impression that you felt offended by my postings. If > my feeling is right, I have to apologize. I never meant to offend you > or anybody else in this newsgroup and I surely won't waste my time on > flame wars. My apologies if I lost my virtual temper. -- Gerry Myerson (gerry@maths.mq.edi.ai) (i -> u for email) Subject: Re: Point on the sphere equidistant from two others Certainly it may be that there are no solutions to this problem. However, when there is, treat the sphere as a constraint and use the sum of the squares of the vectors as the function you're trying to minimize. Use the method of Lagrange multipliers. Example: Let |Z| be a vector. THe sphere constraint is: |Z| - C = 0 The sum of squares function is |A-Z|^2 + |B-Z|^2 Take the gradients of these two vector equations. i.e. take the partial derivatives with respect to each coordinate, the gradient is . Then find if the two gradients are parallel to each other. That will give you the critical points on the sphere. Since your sphere has no boundary in 3-space, any maximum or minimum will be inside the sphere (am I right on this?) Enjoy, -Greg Subject: Re: Point on the sphere equidistant from two others Axel, You can derive an explicit formula for the solution Doing this in cartesian coordinates seems easier to me (I don't remember the spherical coordinates well enough, you might have to tinker with the formulae to accomodate for the usual denominations of N S E W latitude and longitude): let B0 L0 and B1 L1 be your points expressed as latitude and longitude, R the radius of the earth (as R is a common factor in all expressions we will see afterwards, we could as well pose R=1) their cartesian coordinates are x_i = R cos Bi sin Li y_i = R cos Bi cos Li z_i = R sin Bi the great circle of equidistant points belongs to the plane which goes through the center of the earth (origin of the coordinates here) and has the vector going through both points as its normal. Its equation is (x1-x0) x + (y1 - y0) y + (z1 - z0) z =0 if B is the latitude of the parallel you are considering for solutions, the parallel lies on the plane z = R sin B substituting this in the first equations we have the intersection of the two planes (supposing it exists, ie that we don't have x1-x0=0 and y1-y2=0) (x1-x0) x + (y1 - y0) y + (z1-z0) R sin B=0 we are left to find the intersections of this line with the parallel circle of latitude B Let S be the longitude of a point on the parallel, we have x= R cos B cos S y= R cos B sin S replacing this in the line equation, we get the formula for the longitude of the solutions (x1-x0) cos B cos S + (y1 - y0) cos B sin S + (z1-z0) sin B=0 (x1-x0) cos S + (y1 - y0) sin S + (z1-z0) tg B=0 which could be solved (for instance) by using the half tangent equalities posing t= tg (S/2) (beware, there is a special case case when S=Pi) we have cos S = (1- t^2) / (1+t^2) sin S = 2t / (1+t^2) and the equation becomes (x1-x0) (1-t^2) + 2 (y1-y0)t + (z1-z0) tg B (1+t^2) =0 [(z1-z0) tg B - (x1-x0) ] t^2 + 2 (y1 - y0) t + [(z1-z0) tg B + (x1-x0) ] =0 the discriminant is D = (y1 - y0)^2 + (x1-x0)^2 - tg B ^2 (z1 - z0)^2 if it is positive the two solutions are (if all these ugly calculations were correct) S = 2 arctg( -(y1-y0) +/- sqrt(D) ) Finding the closest point is then very easy (it should be the point with the smallest longitude difference, modulo 2pi, to any of the two starting points) Hope this helps Francois 85055d13.0403111233.29125b04@posting.google.com... > I want a formula for the point on an arbitrary parallel of latitude > that is equidistant from two other points on the Earth (assuming a > spherical Earth). So far I used the great-circle-distance expression > to the two given points and equated them. After simplifying there are > two possible identities which can be solved recursively: > Lx = arcsin (( s0 - s1 - (Dc cos Lx) )/Ds ) and > Lx = arccos (( s0 - s1 - (Ds sin Lx) )/Dc ) where > B0,L0 and B1,L1 are the latitude and longitude respectively of the > given points > Bx is an arbitrary latitude > Lx is the longitude of the point sought, Bx,Lx > s0 = sin B0 sin Bx > k0 = cos B0 cos Bx > s1 = sin B1 sin Bx > k1 = cos B1 cos Bx > Dc = ( k1 cos L1 ) - k0 cos L0 > Ds = ( k1 sin L1 ) - k0 sin L0 > I start the recursions with Lx := ( L0 + L1 )/2. > Sometimes one identity works while the other fails to converge. > Sometimes they both fail. > Is there a procedure that always works? > PS - Of course there are two points on the same parallel that are > equidistant to the two given places; I am looking for the closer one. Subject: Re: Point on the sphere equidistant from two others > [ ... ] >Simpler yet (once you're in Cartesian coordinates), add the two >unit vectors and normalize the result. > Did you guys read beyond the original subject line? First off, the set of > points equidistant from two others on a sphere is a great circle. The OP > wants to know where that great circle intersects a given circle of constant > latitude. There are generally 2 such points, and he wants the closer one. Phew! Thank you... Besides, Cartesian coordinates seem very unnatural to me. Subject: Re: Point on the sphere equidistant from two others <4050D590.23DA@mindspring.com> >Do it in cartesian coordinates, and just find the lat and long of the >>midpoint of the straight line between the two points. >Simpler yet (once you're in Cartesian coordinates), add the two >unit vectors and normalize the result. > Did you guys read beyond the original subject line? First off, the set of > points equidistant from two others on a sphere is a great circle. The OP > wants to know where that great circle intersects a given circle of constant > latitude. There are generally 2 such points, and he wants the closer one. > -- spud_demon -at- thundermaker.net There are certainly cases where there is _no_ solution to the OP's problem. Think of the two points lying on the same longitude symmetrically to the equator. Then the great circle that is the location of the points equidistant from both is the equator itsself, which obviously never intersects any other circle of constant latitude away from the equator. Similar for two points with same longitude: The great circle passing through their midpoint will only cover longitudes nearer to the equator, but there will be no solution in the region nearer to the poles than (B0+B1)/2 (assuming the 2 points both on northern or both on southern hemisphere). Also the general case has regions, where there is no solution. Hugo Pfoertner Subject: Re: Point on the sphere equidistant from two others > There are certainly cases where there is _no_ solution to the OP's > problem. Thank you for your comments - however... > Think of the two points lying on the same longitude symmetrically to the > equator. Then the great circle that is the location of the points > equidistant from both is the equator itsself, which obviously never > intersects any other circle of constant latitude away from the equator. The equator is indeed a circle of constant latitude: it just happens to be the only one that is a great circle. In this case the desired point would be Bx=0 and Lx=L0=L1 (in the notation of my first post). Of course you are right to say that the equator would not intersect the locus of points equidistant to the two given points, but if you go back to my OP you will see that I did not state the problem in terms of a locus-intersect. The idea was that given an arbitrary latitude, what is the longitude of the equidistant point? > Similar for two points with same longitude: The great circle passing > through their midpoint will only cover longitudes nearer to the equator, > but there will be no solution in the region nearer to the poles than > (B0+B1)/2 > (assuming the 2 points both on northern or both on southern hemisphere). > Also the general case has regions, where there is no solution. Yes, I should have made the conditions of the problem narrower. But now, what if we specify that the latitude must be equal to (B0 + B1)/2 instead of any arbitrary Bx? Are there still regions with no solution? Of course I am excluding the special case where the two given points are the North and South poles. Subject: Re: Point on the sphere equidistant from two others > There are certainly cases where there is _no_ solution to the OP's > problem. > Thank you for your comments - however... > Think of the two points lying on the same longitude symmetrically to the > equator. Then the great circle that is the location of the points > equidistant from both is the equator itsself, which obviously never > intersects any other circle of constant latitude away from the equator. > The equator is indeed a circle of constant latitude: it just happens > to be the only one that is a great circle. In this case the desired > point would be Bx=0 and Lx=L0=L1 (in the notation of my first post). > Of course you are right to say that the equator would not intersect > the locus of points equidistant to the two given points, but if you go > back to my OP you will see that I did not state the problem in terms > of a locus-intersect. The idea was that given an arbitrary latitude, > what is the longitude of the equidistant point? > Similar for two points with same longitude: The great circle passing > through their midpoint will only cover longitudes nearer to the equator, > but there will be no solution in the region nearer to the poles than > (B0+B1)/2 > (assuming the 2 points both on northern or both on southern hemisphere). > Also the general case has regions, where there is no solution. > Yes, I should have made the conditions of the problem narrower. But > now, what if we specify that the latitude must be equal to (B0 + B1)/2 > instead of any arbitrary Bx? Are there still regions with no solution? No, for this choice the solution can be found by adding the cartesian co-ordinates of the two points and scaling back to r=1 as recommended angles. > Of course I am excluding the special case where the two given points > are the North and South poles. You have to exclude all cases where the two points are exactly opposite to each other on the sphere. In this case the summation of the two vectors gives 0. Any point on the great circle whose plane is perpendicular to both vectors is than a solution to your problem. P0 and P1 at North and South pole is just a special case with any point on the equator being a solution. A more general formulation of the problem with the condition of a given latitude might ask for minimizing the difference of the distances abs(P0-Target) - abs(P1-Target) instead of making those distances equal. This problem will have a solution also in the cases mentioned in my previous posting. Hugo Subject: Trapezoidal Rule & Simpson's Rule - Follow Up Question Thanks for the two examples. Over [-1,1] using n=2 subintervals, the functions f(x)=x^(2/5) and f(x)=cos(pi x) are indeed examples where the Trapezoidal Rule better approximates the definite integral than Simpson's Rule. But for large values of n, Simpson's rule would be the better of the two. Are there situations where the Trapezoidal Rule is better for all n? Thanks again - L Subject: Closure property question Consider a collection of sets F. Define a set C_n of the form oo oo C_n = U / (B_njk), where B_njk belongs to F for all j and k. j=1 k=1 {C_n} is the collection of all such sets. Question: Is {C_n} closed under countable unions? I believe it is since the resulting iterated unions can be expressed as one union, and the resulting set is some C_k. Is this correct, or is there some other reason {C_n} is closed under unions? {C_n} would not necessarily be closed under countable intersections? Is this correct? All help appreciated. Subject: Re: Closure property question > Consider a collection of sets F. Define a set C_n of the form > oo oo > C_n = U / (B_njk), where B_njk belongs to F for all j and k. > j=1 k=1 > > {C_n} is the collection of all such sets. Question: Is {C_n} closed > under countable unions? I believe it is since the resulting iterated > unions can be expressed as one union, and the resulting set is some > C_k. Is this correct, or is there some other reason {C_n} is closed > under unions? > {C_n} would not necessarily be closed under countable intersections? > Is this correct? All help appreciated. Probably I do not understand your question correctly, because this does not make too much sense to me: E_njk is indexed by the three parameters n, j and k. But what does a specific value of n have to do with all the other values of n? Why should the resulting set be some C_k? For example, set E_njk={1,...,n}. Then C_n={1,...,n}. But the union of all C_n is N. The same thing applies for intersections: set E_njk=(N{1,...,n}). Then the intersection of all C_n is empty, though none of the C_n is. -- I'm on warm milk and laxatives Cherry-flavored antacids reverse my forename for mail! - saibot Subject: Re: Strange Complex Variables Problem > Oh my gosh, no I didn't say anything about the residue. > I said something about the coefficient of 1/z in the Laurent > expansion, but (i) that was when I was talking about > functions in an _annulus_, not in D, (ii) that coefficient > is _not_ a residue (except in the case when the annulus > has zero inner radius.) Yeah, that's true. The 2pi might not be the result of a residue, unless f is meromorphic in the region bounded by the outer curve. I didn't realize what you were saying. That's right, you can't even always get a laurent series expansion in my arbitrary domain D. Well you can but for small circles within D. I assumed you were talking about residues as f went around D's inner curve because it has a pole inside the inner curve. No, basically the i*2PI is there to prop up the functions so their L2 norm could be bounded below by F'/F . -Greg Subject: Re: Strange Complex Variables Problem I also wanted to mention again that in addition to what I said, it's useful to know that square-integrable analytic functions DO form a hilbert space on the domain, because these functions converge in the L2 norm iff they converge uniformly (i.e. to another analytic function). -Greg Subject: Re: Strange Complex Variables Problem >I also wanted to mention again that in addition to what I said, it's >useful to know that square-integrable analytic functions DO form a >hilbert space on the domain, because these functions converge in the >L2 norm iff they converge uniformly (i.e. to another analytic >function). That's nonsense. Uniform convergence does not imply convergence in your Hilbert space, unless D has finite area. And convergence in your Hilbert space does not imply uniform convergence, regardless of what D is. The reason that that space is a Hilbert space, now that we've finally got straight what space we're talking about, is that convergence in that space implies uniform convergence _on_ compact subsets of D (hence that the limit is analytic.) >-Greg ************************ Subject: Re: Strange Complex Variables Problem >Well, I'm talking about the L2 norm induced by the inner product = INTEGRAL(over D) [ f(z) conj[g(z)] ] dz , which is simply L2(f) = >INTEGRAL(over D) [ |f(z)|^2 dz ]. > This is (in my experience) a strange use of notation. I meant to indeed say INTEGRAL (over D) [ f(z) g(z)* ] dz dz* [ I meant to say Green's theorem to integrate around the boundary but that's irrelevant. Let me try to start over because I don't want to confuse anyone. ] I've got a hilbert space H of analytic functions defined on a region D, square-integrable on D. i.e. Integral [over D] ( |f|^2 dz dz* ) or alternatively Integral [over D] ( |f|^2 dx dy ) exists and is finite. Take the L2 norm of f as the square root of this integral, and call it L2(f). The region D is doubly-connected. There is a function F conformally mapping the region D into an annulus centered around the origin. The radii of this annulus are not specified -- it's explained that because this is a conformal map, their ratio is always the same and F'/F sensitive only to their ratio. (e.g. G = 2F, G'/G = F'/F, etc.) Now take the subset J of H consisting of functions with the following property: the function integrated along any curve going around the hole in the positive direction has a residue of 1. i.e. it has the value i*2pi . F'/F is certainly square-integrable. How can I show that L2(F'/F) <= L2(j) for any j in J? I hope that cleared things up. Sorry if I was misstating the problem. -Greg PS: Yes, I am using standard lebesgue measure here to integrate, I believe you can even use the regular Riemann integral since we are always integrating analytic functions, which uses a measure which mathworld calls with the funny name of Jordan measure. Subject: Re: Strange Complex Variables Problem >>Well, I'm talking about the L2 norm induced by the inner product >= INTEGRAL(over D) [ f(z) conj[g(z)] ] dz , which is simply L2(f) = >>INTEGRAL(over D) [ |f(z)|^2 dz ]. > >> This is (in my experience) a strange use of notation. >I meant to indeed say INTEGRAL (over D) [ f(z) g(z)* ] dz dz* >[ I meant to say Green's theorem to integrate around the boundary but >that's irrelevant. Let me try to start over because I don't want to >confuse anyone. ] >I've got a hilbert space H of analytic functions defined on a region >D, square-integrable on D. i.e. Integral [over D] ( |f|^2 dz dz* ) or >alternatively Integral [over D] ( |f|^2 dx dy ) exists and is finite. >Take the L2 norm of f as the square root of this integral, and call it >L2(f). >The region D is doubly-connected. There is a function F conformally >mapping the region D into an annulus centered around the origin. The >radii of this annulus are not specified -- it's explained that because >this is a conformal map, their ratio is always the same and F'/F >sensitive only to their ratio. (e.g. G = 2F, G'/G = F'/F, etc.) >Now take the subset J of H consisting of functions with the following >property: the function integrated along any curve going around the >hole in the positive direction has a residue of 1. i.e. it has the >value i*2pi . >F'/F is certainly square-integrable. No, it's not true that F'/F is certainly square-integrable! I explained this a few days ago. Suppose that D is the unit disk minus the origin. Then D is certainly doubly-connected, and your annulus is D itself; F(z) = z, and then F'/F is _not_ square-integrable over D. Answer me a question. Is this a problem like in a book somewhere? I know that you know the answer to _that_ question... if it is a problem in the book could you tell us _exactly_ how the problem reads? >How can I show that L2(F'/F) <= >L2(j) for any j in J? >I hope that cleared things up. Sorry if I was misstating the problem. >-Greg >PS: Yes, I am using standard lebesgue measure here to integrate, I >believe you can even use the regular Riemann integral since we are >always integrating analytic functions, which uses a measure which >mathworld calls with the funny name of Jordan measure. ************************ Subject: Re: all the roots of a nonlinear equation > I realize this is probably a dumb question but I've looked at numerical > analysis books and racked my brain but I must be missing something obvious. > I have equations which are sums of sines and cosines an example being: > 2 Cos[x] - Cos[2 x] - Sin[x] + 3 Sin[2 x] - Sin[3 x] = 0 > I need to find all the roots of this equation in the range -pi to pi. > The algorithms I've looked at only give me one root and if based on a > bisection technique dont always give an answer in the range I mentioned > above. > I tried using NSolve in mathematica for this above equation and it does'nt > return any roots (there are 4 when looking at the plot). > I am interested in an algorithm that I could implement in C / Fortran for > later use that would give me all the roots of the above types of functions > automatically. > Is this possible? Or have I missed something very obvious? > Thanks, Would the following work? Write cos(nx) as (t^n + t^-n)/2, write sin(nx) as (t^n - t^-n)/2i, clear denominators, and find all of the complex roots of the resulting polynomial in t. Then you know sin(x) and cos(x), so easy to find x. So your sample equation 2 Cos[x] - Cos[2 x] - Sin[x] + 3 Sin[2 x] - Sin[3 x] = 0 becomes (t+t^-1) - (t^2+t^-2)/2 - (t-t^-1)/2i + 3(t^2-t^-2)/2i - (t^3-t^-3)/2i = 0. Multiplying ty 2it^3 gives 1 - 3t + (-1+2i)t^4 + (1+2i)t^2 + (3 - i)t^5 - t^6 = 0. The roots of this (according to PARI) are -1.09973 + 0.261803550*I -0.025318021 - 2.89325*I 0.233951633 - 0.649323431*I -0.040888745 - 1.44359*I 0.938942371 - 0.047583210*I -0.006953501 + 1.63036*I and the corresponding [cos(x),sin(x)] are [0.321614201, 0.086177678] [-0.945089837, -0.239647076] [1.00643, -0.764022744] [0.121776434, -0.952180855] [2.55438, -0.121637430] [-0.059117022, 0.991309694] Joe Silverman Subject: Re: all the roots of a nonlinear equation JHS escribi.97 en el mensaje >> I realize this is probably a dumb question but I've looked at >> numerical >> analysis books and racked my brain but I must be missing something >> obvious. >> I have equations which are sums of sines and cosines an example >> being: >> 2 Cos[x] - Cos[2 x] - Sin[x] + 3 Sin[2 x] - Sin[3 x] = 0 >> I need to find all the roots of this equation in the range -pi to pi. >> The algorithms I've looked at only give me one root and if based on a >> bisection technique dont always give an answer in the range I >> mentioned >> above. >> I tried using NSolve in mathematica for this above equation and it >> does'nt >> return any roots (there are 4 when looking at the plot). >> I am interested in an algorithm that I could implement in C / >> Fortran for >> later use that would give me all the roots of the above types of >> functions >> automatically. >> Is this possible? Or have I missed something very obvious? >> Thanks, > Would the following work? Write cos(nx) as (t^n + t^-n)/2, write > sin(nx) as (t^n - t^-n)/2i, clear denominators, and find all of the > complex roots of the resulting polynomial in t. Then you know sin(x) > and cos(x), so easy to find x. > So your sample equation > 2 Cos[x] - Cos[2 x] - Sin[x] + 3 Sin[2 x] - Sin[3 x] = 0 > becomes > (t+t^-1) - (t^2+t^-2)/2 - (t-t^-1)/2i + 3(t^2-t^-2)/2i - (t^3-t^-3)/2i > = 0. > Multiplying ty 2it^3 gives > 1 - 3t + (-1+2i)t^4 + (1+2i)t^2 + (3 - i)t^5 - t^6 = 0. I get - t^6 + t^5(3 - i) + t^4(2i - 1) + t^2(2i + 1) - t(i + 3) + 1 = 0 > The roots of this (according to PARI) are > -1.09973 + 0.261803550*I > -0.025318021 - 2.89325*I > 0.233951633 - 0.649323431*I > -0.040888745 - 1.44359*I > 0.938942371 - 0.047583210*I > -0.006953501 + 1.63036*I > and the corresponding [cos(x),sin(x)] are > [0.321614201, 0.086177678] > [-0.945089837, -0.239647076] > [1.00643, -0.764022744] > [0.121776434, -0.952180855] > [2.55438, -0.121637430] > [-0.059117022, 0.991309694] > Joe Silverman But two of that solutions aren't in the (real) interval (-pi, pi] ... -- Best regards, Ignacio Larrosa Ca.96estro A Coru.96a (Espa.96a) ilarrosaQUITARMAYUSCULAS@mundo-r.com Subject: Re: Are you able? > A pproblem in http://santinho-de-pau-a-fala.blogspot.com/ > Manuel Sousa, Portugal The formula W/10000 = (1-1/n)^n gives 3678,77601766571 for n=100000. Subject: Re: Is this by support1.mathforum.org (8.11.6/8.11.6/The Math Forum, $Revision: 1.9 primary) id i2CGkI102949; The open path problem, in the case where all cities must be visited, is the Hamiltonian path problem, which is also NP-hard. >> If you are required to return to the starting point, >> then this is the Hamiltonian cycle problem, >> which is knownto be NP-hard. >No. The path is allowed to be open (and in the problem I'm working on, >it is always open except we have only one city) >I din't even see a proof of NP-hardness for the open path version of >TSP. Here help is appreciated too. >-- >Best regards, >Alex. >PS. To email me, remove loeschedies from the email address given. Subject: Re: Cantor's Diagonal Argument In sci.logic, |-|erc : >> Similarly : >> Prob(exists j, for i = 1..n, b_i = c(j)_i)=1 -> Prob(exists j, >> for i = 1..n+1, b_i = c(j)_i)=1 >> >> By induction on n : >> for all n, exists j, b_1 = c(j)_1, b_2 = c(j)_2, ... b_n = c(j)_n >> >> This suggests NotEqual(b, c(j)) will not halt. >> It won't halt anyway; that's part of the problem. >> Assume one is testing all b's. A UTM that enumerates >> all b's can be created (the problem then becomes >> slightly recursive!) and one can then run NotEqual(b, c(j)) >> on all j, while running simultaneously the b-generator UTM >> and doing the comparisons with oodles of threads. >> A pretty and complex problem -- but this problem will >> not halt regardless, as there are an infinite number >> of b's to test. > That's not a concern. The initial formulation of Cantor's proof is that it > finds ONE b! We take that number and however many digits we compare > its computable. Its effective to demonstrate a proof that shows that > the 1 formulated number is on the list. b is only computable if the list of numbers is computable, which one has to assume if only because the mapping c(j) is specified. > Or as you originally put it : >>Your problem is slightly different; you need to show that for *all* b > which is actually where my proof begins, covering the extended problem > of multiple b's. > For a decimal matrix there are 8 to 9 possibilites for each digit of b, > the number of b's is just 9^digits to test. Parallel testing is no problem, > but not necessary, if the Cantonian formulation is invalid for one of the > contenders the others will likely fail too. Go for it. >> It won't halt anyway; that's part of the problem. > This is only a problem for multiple c(j), we can consider b a constant. > //Globals > b = Cantor(C, random) > cj = FindMatch(C, b, 1) > DiagNotEqual () { > DigitsEqual (b, cj, 1) > Return TRUE > } > DigitsEqual (DE1, DE2, digit) { > if (DE1_digit = DE2_digit) { > DigitsEqual (DE1, FindMatch(C, b, digit+1), digit+1) > } > } > Here FindMatch will tally through computable numbers until it finds > a number that matches the initial digits of b. > j is allowed change to a different number during testing. This algorithm > will still determine whether a list of numbers does not contain a new > test number. The algorithm will not halt on the list of computables against > a new number however it is constructed. > Unfortunately the algorithm is too giving, it would never determine that pi > was off the list of rationals, it would keep constructing larger fractions to > continue the comparison. > To me its enough that any number off the list is nonsensical, > it only works when that number is infinitely long, which voids > it from any mathematical construction. It doesn't work anyway as an algorithm. It only works as a semialgorithm; the routine halts and announces failure, but goes into an infinite loop upon success. But that's not a problem, mathematically. All you have to do is prove that DiagNotEqual(b, c) never halts for a Cantorian-constructed b. > Herc -- #191, ewill3@earthlink.net It's still legal to go .sigless. Subject: Re: Cantor's Diagonal Argument oo ____|mn / /_/ / _ / K-9/ /_/ - www.YeOldeCoffeeShoppe.com - /____/_____ -------------- >> Prob(exists j, for i = 1..n, b_i = c(j)_i)=1 -> Prob(exists j, >> for i = 1..n+1, b_i = c(j)_i)=1 >> >> By induction on n : >> for all n, exists j, b_1 = c(j)_1, b_2 = c(j)_2, ... b_n = c(j)_n >> >> This suggests NotEqual(b, c(j)) will not halt. > >> It won't halt anyway; that's part of the problem. > >> Assume one is testing all b's. A UTM that enumerates >> all b's can be created (the problem then becomes >> slightly recursive!) and one can then run NotEqual(b, c(j)) >> on all j, while running simultaneously the b-generator UTM >> and doing the comparisons with oodles of threads. > >> A pretty and complex problem -- but this problem will >> not halt regardless, as there are an infinite number >> of b's to test. > > That's not a concern. The initial formulation of Cantor's proof is that it > finds ONE b! We take that number and however many digits we compare > its computable. Its effective to demonstrate a proof that shows that > the 1 formulated number is on the list. > b is only computable if the list of numbers is computable, which > one has to assume if only because the mapping c(j) is specified. Ouch, I thought I defined the mapping given the usage of a UTM. See It was a long procedure I outlined, basically to overcome non halting programs in the list of all functions' outputs. Say all programs halt, then UTM(1) = some number, UTM(2) = some number, UTM(3) = some number. The set of numbers output from UTM applied to N in turn, contains all computable numbers. All I did was list the ones that halt 1st, testing digit by digit, to guarantee that a digit is present on the diagonal. > Or as you originally put it : >>Your problem is slightly different; you need to show that for *all* b > which is actually where my proof begins, covering the extended problem > of multiple b's. > For a decimal matrix there are 8 to 9 possibilites for each digit of b, > the number of b's is just 9^digits to test. Parallel testing is no problem, > but not necessary, if the Cantonian formulation is invalid for one of the > contenders the others will likely fail too. > Go for it. Ouch!! Doesn't this mean b = c(j)? and is it correctly derived? for all n, exists j, b_1 = c(j)_1, b_2 = c(j)_2, ... b_n = c(j)_n This is equivalent to : for all n, b_n = c(j)_n >> It won't halt anyway; that's part of the problem. > This is only a problem for multiple c(j), we can consider b a constant. > //Globals > b = Cantor(C, random) > cj = FindMatch(C, b, 1) > DiagNotEqual () { > DigitsEqual (b, cj, 1) > Return TRUE > } > DigitsEqual (DE1, DE2, digit) { > if (DE1_digit = DE2_digit) { > DigitsEqual (DE1, FindMatch(C, b, digit+1), digit+1) > } > } > Here FindMatch will tally through computable numbers until it finds > a number that matches the initial digits of b. > j is allowed change to a different number during testing. This algorithm > will still determine whether a list of numbers does not contain a new > test number. The algorithm will not halt on the list of computables against > a new number however it is constructed. > Unfortunately the algorithm is too giving, it would never determine that pi > was off the list of rationals, it would keep constructing larger fractions to > continue the comparison. > To me its enough that any number off the list is nonsensical, > it only works when that number is infinitely long, which voids > it from any mathematical construction. > It doesn't work anyway as an algorithm. It only works as a semialgorithm; the routine halts and announces failure, but > goes into an infinite loop upon success. sprung, I was lazy with the construction. FindMatch goes to the next number if there is no match, fails the next digit comparison, falls out of the loop as not equal. I should write a general function that works with any list and a number. > But that's not a problem, mathematically. All you have to do > is prove that DiagNotEqual(b, c) never halts for a > Cantorian-constructed b. I think I came close, but if it proves diag is computable it also proves pi is rational. I'd had to overhaul my assumptions at this point, work 'can be calculated' into my defn of computable. Herc Subject: algebra: additive,abelian I am learning about additive, abelian category. The book quote an example: Let C be a catgeory such that the object of C are pairs (A',A) where A is an abelian group and A' is a subgroup of A. Morphism from (A',A) to (B',B) is a group homomorphism f: A to B where f(A') is in B'. The book claim that it is an additive category having kernel and cokernel of each morphism but it is not abelian. Why is it true? Can anyone tell me how to see it? Subject: Re: algebra: additive,abelian Adjunct Assistant Professor at the University of Montana. > I am learning about additive, abelian category. The book quote an >example: >Let C be a catgeory such that the object of C are pairs (A',A) where A is an >abelian group and A' is a subgroup of A. Morphism from (A',A) to (B',B) is >a group homomorphism f: A to B where f(A') is in B'. >The book claim that it is an additive category having kernel and cokernel of >each morphism but it is not abelian. >Why is it true? Can anyone tell me how to see it? To see it is additive, you need the homomorphism sets to be an additive abelian group in which composition is bilinear, which will be inherited from the category of Abelian groups; and a zero object, and a biproduct. The biproduct is just the direct sum on each coordinate: (A',A) oplus (B',B) = (A'oplus B', A oplus B) which can be seen to have the necessary universal properties. The zero object is the pair (0,0). The kernels should be the usual kernels: the kernel of f:(A',A)->(B',B) will be (ker(f) intersect A', ker (f)). The real difficulty is the cokernels. One might be tempted to define coker(f:(A',A)->(B',B)) = (B'/f(A'), B/f(A)) but that definition does not work because in general B'/f(A') is not a subgroup (or isomorphic to a subgroup) of B/f(A). Rather, you need to take [B'+f(A')]/f(A) = B'/[f(A) intersect B']. So we define; coker(f:(A',A) -> (B',B) ) = ([B'+f(A')]/f(A) , B/f(A)) To see it is a cokernel, we need to show that it is the coequalizer of f and the zero map from (A',A) to (B',B). That is, we need to show that coker(f)f = 0 [which it is], and that if h:(B',B) -> (C',C) has the property that hf = 0, then h = h'coker(f) for some unique arrow h'. If hf = 0, then f(A) is contained in the kernel of h, so the map B->C factors uniquely through B/f(A), which gives you the property. The usual universal properties of the kernel and cokernel in Ab will give you the necessary properties here. Now, an abelian category is an Ab-category which has a zero object, binary biproducts, every arrow has a kernel and a cokernel, and every monic is a kernel and every epi a cokernel. So in this case, what you want to show is that not every monic is a kernel, or not every epi is a cokernel (since the category satisfies all the others). The key is that the cokernel is not (B'/f(A'), B/A). This is what is going to fail. So we can find an epi which is not a cokernel, say by finding a map which is an epi from A to B, but whose restriction from A' to B' is not. So we look for a map which is surjective A->B, but where f(A) intersect B' is larger than f(A'); say, by letting B'=B, but A' smaller than A which does not map onto B' and is not contained in the kernel of f. Consider A' = Z/2Zx{0}, A= (Z/2Z)x(Z/4Z), B' = B = Z/4Z, and the map f(a,b) = 2a+b. I claim this is an epi. For if g,h:(B',B)->(C',C) satisfy gf = hf, then g(1) = g(f(1))=h(f(1))=h(1), so g=h, showing that f is right cancellable. But the thing is that f is not a cokernel. For assume that there is a map g:(C',C) -> (A',A) so that f is the coequalizer of g and the zero map 0:(C',C)->(A',A). Then that means that gf=0. So g(C) is contained in {(0,0), (1,2)}. In either case, g(C') = 0. g cannot be the zero map, because the coequalizer of the zero map with itself is the identity. So g must map C onto {(0,0), (1,2)} and C' to 0. Let N = {(0,0),(1,2)}. If h:(A',A) -> (D',D) is any map such that hg = 0, then it should factor through f. But take the map h:(A',A)->([A'+N]/N,A/N}) given by the cacnonical projection. Then hg=0, so the map should factor through f. That is, there exists h':(B',B) -> ([A'+N]/N, A/N) so that h = h'f. Since B'=B, the image of h' must be contained in [A'+N]/N = {(a,b): a in Z/2Z, b=0 or 2} But what is h'(1)? h'(1) = h'(f(0,1)) = h(0,1) = (0, 1+N), which is not in [A'+N]/N. So the map cannot factor through f, so h cannot be a cokernel. -- ============================================================== ======== It's not denial. I'm just very selective about what I accept as reality. --- Calvin (Calvin and Hobbes) ============================================================== ======== Arturo Magidin magidin@math.berkeley.edu Subject: sine series by support1.mathforum.org (8.11.6/8.11.6/The Math Forum, $Revision: 1.9 primary) id i2CHxs611733; The series sin x = x - (x^3)/3! + (x^5)/5! - ... doesnot works when x is large, eg. x = 80 Is there any series which is generalized for computing sine for any x? Subject: Re: sine series > The series > sin x = x - (x^3)/3! + (x^5)/5! - ... > doesnot works when x is large, eg. x = 80 > Is there any series which is generalized for computing sine for any x? It does work, but it will not begin to converge before the number of terms is much larger than x (before, successive terms increase...). If x is very large, you need to calculate a lot of terms before you get convergence, and you might get an overflow of roundoff errors. To be able to effectively use this series, substract an appropriate multiple of 2 PI from x so that y = x - 2 k PI is between -PI and PI and calculate sin(x) using y in the formula above. (you can do a bit better by reducing the argument to -PI/2 PI/2 and using the symetry of the sine function around PI/2) Note also that if you want to calculate an approximation to the sine function using a polynomial of fixed degree, the above formula is not the best approximation. Francois provided you first reduce x to a smaller number, : noting that sin (x+2kPI) = sin(x) for any x and integer k, substract a multiple of 2kPI from x, so that Subject: Re: sine series > The series > sin x = x - (x^3)/3! + (x^5)/5! - ... > Note also that if you want to calculate an approximation to the sine > function using a polynomial of fixed degree, the above formula is not the > best approximation. Is a Hermite interpolation the best way of implementing sine on a computer? I know the name, but haven't figured it out yet. -- Try http://csf.colorado.edu/pkt/pktauthors/Vienneau.Robert/ Bukharin.html To solve Linear Programs: .../LPSolver.html r c A game: .../Keynes.html v s a Whether strength of body or of mind, or wisdom, or i m p virtue, are found in proportion to the power or wealth e a e of a man is a question fit perhaps to be discussed by n e . slaves in the hearing of their masters, but highly @ r c m unbecoming to reasonable and free men in search of d o the truth. -- Rousseau Subject: Re: sine series > The series > sin x = x - (x^3)/3! + (x^5)/5! - ... > doesnot works when x is large, eg. x = 80 Depends on what you mean by work. The radius of convergence is infinite, and therefore the series is mathematically correct for all real x. However, machine computation with limited precision may fail when x is large. > Is there any series which is generalized for computing sine for any x? You mean, you want one series that works for all x without significant roundoff error? Not likely. Instead of the MacLaurin series, you could try using the Taylor series expansion about an arbitrary x0. In the case of the sin function there is a particularly simple series when x0 = 2*pi*n: sin(x-2*pi*n) = (x-2*pi*n) - (x-2*pi*n)^3/3! + (x-2*pi*n)^5/5! - ... This works for values of x that are reasonably close to 2*pi*n. Of course, it's effectively the same as just performing a trigonometric reduction on x and then applying the MacLaurin series to the reduced argument. -- Dave Seaman Judge Yohn's mistakes revealed in Mumia Abu-Jamal ruling. >Here's a new game: player 1 picks a number that may or may not >>be transcendental, player two wins if he can prove whether the >>number is transcendental or not. >>I think player 1 will win almost every time, but if player 2 wins, >>maybe he'll get a Nobel prize! >Only if he can publish the proof in an economics journal. Oops! Forgot about Mr. Nobel's grudge against mathematicians. George >Robert Israel israel@math.ubc.ca >Department of Mathematics http://www.math.ubc.ca/~israel >University of British Columbia >Vancouver, BC, Canada V6T 1Z2 Subject: Matrix norms. Is it possible to define an unbounded norm on matrices? Subject: Re: Matrix norms. >Is it possible to define an unbounded norm on matrices? You must not be asking exactly the question you meant to ask: _every_ matrix norm is unbounded! (Proof: If ||.|| is a matrix norm and A <> 0 then ||A|| > 0. But ||cA|| = c||A|| for c > 0, and c||A|| is unbounded.) Trying to guess maybe what you meant: any norm is _finite_, by definition. ************************ Subject: Re: Matrix norms. > Is it possible to define an unbounded norm on matrices? Perhaps that I am missing something here, but the only vector that I know in which you can define a bounded norm is {0}. Best regards, Jose Carlos Santos Subject: Unit pricing: Open-ended mutual funds The value of individual units in open-ended mutual funds / unit trusts are given by the formula: unit value = value of investment fund / number of units in issue to clients An adjustment is made for spread to determine the sale and repurchase price of the units. The value of the units increases when the value of the fund increases (and visa versa) The aim of the mutual fund company is therefore to increase the value of the investment fund (portfolio) in order to achieve an exceptable return for investors. The emphasis is therefore on increasing the numerator in the equation. Gain Theory suggests that aditional gain can be achieved by also limiting the denominator in the equation, i.e. the number of units that are issued to clients. This can be easily achieved by capitilising the spread into the value of the investment fund. For more info refer to: http://www.verismall.com/software/other/unittrustpricingmodel/ unittrustprici ngmodel.htm Francois Terblanche Verismall Software www.verismall.com Subject: Conencted Small Graphs Originator: fab@soda.csua.berkeley.edu (Fabio Rojas) Not sure if this was posted the first time... Is there a reference that tells me the % of connected graphs with 30 nodes or less? How about equivalence classes of graphs? Formula would be nice, if it exists. The table would look like: Nodes Percent Connected 3 50% ..... 30 .000003% Feel free to email me or post a citation. Fabio Subject: Accounting of downloads The second volume of Buckminster Fuller's books Synergetics was published in 1979, twenty five years ago. All of humanity had to learn things like the synergetics coordinate system to do more with less soon or it would be curtains for humanity according to Bucky, but look at these low numbers. And life still goes on. Cliff Nelson Begin forwarded message: Dear Clifford Nelson, At the moment, I only have access to the top 150 downloads. You have two notebooks in the top 150. (I only have one, myself). --Ed Pegg Jr. Could you give me an accounting of the number of downloads of my notebooks from MathSource please? http://library.wolfram.com/infocenter/search/?search_results=1 ;search_person _id=607 ? Buckminster Fuller Notebooks [in /MathSource: Packages and Programs/] #129 in all time downloads. In the past year, notebook downloaded 1118 times ? Bucky Number Mandelbrot [in /MathSource: Packages and Programs/] #86 in all time downloads. In the past year, notebook downloaded 1335 ? Four Triangle Fractals using Bucky Numbers and Synergetics Coordinates [in /MathSource: Packages and Programs/] Not in top 150 ? Solving Matrix Problems Using Bucky Numbers [in /MathSource: Packages and Programs/] Not in top 150 Cliff Nelson Subject: Re: Catholic or Protestant background ? charset=iso-8859-1 > Hi Ha Ha Hanson, You said that you're not Jewish. > Do you come from a Catholic or a Protestant background ? > Do you believe in a Benevolent Creator ? Your need to know is truly touching, Professor Relf. Therefore, we are supremely honored to fulfill your splendid inquisitorial nosiness, in great detail, in the CV that follows. We all do hope that the answers provided therin will ameliorate your anxieties that originate from your vacuous curiosity. First of all, hanson@quick.net is a cyber construct, a composite from dozens of contributors who post under this handle, to this & other such cyber parties aka News-groups. Most of the posts do originate after board/court/R&D/prod biz meetings to **vent**, or out of **boredom** from up at some 35-45kft during the daily 17 hrs flight durations, or from-to some godforsaken locations, at/in/to anywhere, to **reassure** our dudes/dudettes, etc, etc, & of cousre, etc.... hanson@quick.net should be viewed from the perspective that it has a reality equivalent/content like the following cyber entity, which you should study intensively, in particular its activities at Ravencrag : http://www.marvunapp.com/Appendix/cobrajal.htm#Hanson http://www.marvunapp.com/Appendix/cobrajal.htm#Ravencrag Now, to the details of your requests, Professor Relf: > Hi Ha Ha Hanson, You said that you're not Jewish. There is a general uncertainty about that, and as general Jew propaganda likes to have it, I have no disprove that some rabbi, at one time, did not beat off into the family tree of my European ancestors. Stranger things have happened in the days of yore, whore, yondern that makes you wondern. --- Hallelujah, praise the Load! --- To boot, I am still looking for a Jewish Ersatz Grandpa. One like that would come in handy, if you wish to make points with fanatical Jews. I scoured Yiddishe olde kacker homes for an Ersatz Granpa, for weeks, but they all wanted a big time fee for it, upfront, even before signing on, with no guarantee for standing up for me when it would count, cyber or otherwise. So, go figure, as they say, ...and oy vey...... > Do you come from a Catholic or a Protestant background ? They used to send me to some ing Sunday Shul, were the old, holy looking broad was distributing pretty postage stamps with Jesus+halo on it so that we kiddies would keep quiet. I have no idea what this crock was all about. Then, during my 1st high intensity testosterone season, the religo/spiritual establishment enslaved me again into another gig of theirs, attempting boring protesta/barmitz/commun type brainwashing, which was so full of transparent lies and crock that was read out of an old fat book and a roller that looked like a 17th century toilet paper roll. .....And then these wailing, ing songs!........Jesus-H-ing Kriste ...........depressing. If it would have been the Beatles or Elvis....well........ So anyway, there was one good thing about all this, ....the wife of the preacher was some good looking, pretty, beautiful ass- and boob wiggling bitch. Every time she looked at me I got an instant hard on. And she knew it and moaned while passing by me. ...But the day of rapture did come! The Load was smiling upon me then. Passion overcame beautiful Mrs. Preacher and she grabbed & pushed me onto the altar and gave me one of the great and memorable rides......Just as she hollered and her juices ran down over my crotch the preacherman came in and asked what we where doing.....---ex-communication, etc.. Went to a bunch of schools, the military, behaved, (well, maybe) till, still drunk, right after the doctoration ceremonies, my ex-high school cronies at the bar told me that I would never catch up with their earnings until I would be 55, if lucky....... Having just gone thru 4 years of penniless academic slavery, the realization of this negative $$-aspect-future made me so furiously mad that marched right away over to my now ex-professor/advisor's office where I placed my cum laude diploma carefully into my square tassel hat, took a onto it and poured a half a quart of good vodka over it. Threw a match at it and watched the beautiful faint blue flame evolve into a yellow sodium fire that began to stink more & more. Just as I was celebrating my achievement professor schmuckface, as I called him, came waltzing in and asked me what I was doing..... deja vue........---ex-communication, etc....... left town in a jiffy. Never looked back .......went to see the world.......ahahahahaha....... So, as you can see, professor Relf, they did drag you to Brazil to proselytize FOR somebody.....I, OTOH, dragged my own ass also to far away stray corners of the globe, BUT I proselytized for myself and for my own cause......biz........ Moral of the story, both of us had a very banal youth, didn't we. Such then, this question of yours did not furnish you anything new. > Do you believe in a Benevolent Creator ? Why, Professor Relf? Can't you, you of all people, can't you see that I am one of them. -- May I administer to you, my son? What are your spiritual needs, today, son Jeff? However, you must know that there are my personal permit charges and user fees attached to such services. Payable up front. Like the govt's. envrionmental tax levies.....the creation is green, after all.... However, on a deeper, more useless level I can answer your question for free, but with equal benevolence: Everything, all actions or any person that bestows benevolence upon you is a Benevolent Creator. You know, Professor Relf, there are some very nice people and many good opportunities around that are benevolent, even if they are not Jewish. Or even more general, a case can be made that everything which can kick your ass, take away your control, determine your behavior etc. is a benevolent creatoration. With benevolent being a term that is higlhy relative when seen from and in different frames....... However, if the Benevolent Creator must take on some mystical dimension for your existential salvation validification then I do recommend for you to say: Hallelujah, Praise the Load!, Professor Relf ahahahahaha.........AHAHahahhahaha...... ahahahahanson Subject: Harvard.EDU . Hi Ha Ha Hanson, You called me Professor Relf . Actually, I'm the president of Harvard University ... http://www.president.Harvard.EDU/ . . . but I'll let that slide. You said, With ' benevolent ' being a term that is highly relative when seen from and in different frames . . . . Ah, so you are a theoretical physicist then. Stevie Hawking perhaps ? Subject: Re: Three positive integer triples with equal sums and products > I am looking at directions for solving the following problem, > Given a set of three strictly positive integers {a,b,c} (a>0,b>0,c>0) > Under which conditions can I find two other sets of three positive integers > {d,e,f}and {g,h,i}, with the same sum and product > ie > a+b+c=d+e+f=g+h+i > abc=def=ghi > with sets {a,b,c} {d,e,f} and {g,h,i} all different (not permutations of the > same three numbers) > (this could be generalised to 4, 5 ... triples, ad nauseam) > I have been working on this for a while, and could prove a few basic > theorems on these sets (eg we cannot have a=d, neither can we have a=b=1, > nor a=b=c, etc...), but could not find a constructive way of building either > all or some of the solutions to this problem. Do you allow duplicates within a triplet as long as the triplets themselves are unique? The following three triplets have a sum of 49 and product of 3600 12 25 12 10 15 24 9 20 20 Or must all 9 numbers be unique, as in the following sum of 49, product of 3024? 9 12 28 8 14 27 7 18 24 > Does anyone have an idea on how to deal with this? I generated mine using an Access database, but that's probably not what you want. > In case someone is interested, it came as a math puzzle, about three > children, receiving three presents in three different boxes, all of which > had ineger dimensions, and noticing that the three boxes had the same > volume, and that the pieces of string packing them had the same lengths...) > Thanks in advance > Francois Subject: Re: Three positive integer triples with equal sums and products > Do you allow duplicates within a triplet as long as the triplets > themselves are unique? The following three triplets have a sum of > 49 and product of 3600 > 12 25 12 > 10 15 24 > 9 20 20 yes, so far the three triplets are different, it is ok. It is possible to prove, though - that there cannot be a solution with the same integer in two different triplets (else, the conditions would become a+b=d+e and ab=de, which mean the two triplets are equal) - that in no solution can a triple be a cube (a a a) - that although you can have one or two triplets with two duplicate integers, as in your example, you cannot have three triplets with two duplicate integers in one solution Francois Subject: Re: Three positive integer triples with equal sums and products > I am looking at directions for solving the following problem, > Given a set of three strictly positive integers {a,b,c} (a>0,b>0,c>0) > Under which conditions can I find two other sets of three positive integers > {d,e,f}and {g,h,i}, with the same sum and product > ie > a+b+c=d+e+f=g+h+i > abc=def=ghi > with sets {a,b,c} {d,e,f} and {g,h,i} all different (not permutations of the > same three numbers) > (this could be generalised to 4, 5 ... triples, ad nauseam) > I have been working on this for a while, and could prove a few basic > theorems on these sets (eg we cannot have a=d, neither can we have a=b=1, > nor a=b=c, etc...), but could not find a constructive way of building either > all or some of the solutions to this problem. > Does anyone have an idea on how to deal with this? This is discussed in problem D16 in Guy, Unsolved Problems in Number Theory. The 3rd edition is in press. The 2nd edition says Schinzel has proved that you can have as many triples as you like of positive integers with the same sum and the same product, but it doesn't give a citation. It does give some other information that might interest you, and references to Problem E2872 of the American Math Monthly, 89 (1982) 499-500 and 88 (1981) 148. -- Gerry Myerson (gerry@maths.mq.edi.ai) (i -> u for email) Subject: Re: Sound levels charset=iso-8859-1 > It seems to me that if there are x machines each with a sound level > of L(1) then the level of x machines should be > L(x)=x{10log[I(1)/I(0)]} > rather than what was given to me as above. > I don't even know if I'm missing the concept or the maths! The concept. A level *is* the logarithm of a ratio of whatever. If multiply something with 10, you could as well call it a 10 dB increase, if you multiply something with 100, you call it a 20 dB increase, multiplying with 1000 gives 30 dB and so on. And when something is twice as much than something else, you could say it's got 3 dB more, because 10 lg 2 = 3 (well, more or less) > Please explain as if to a 5 year old *smiles* ... this one has really > got me confused! I like this. :-) You can now inmprss your friends in the kindergarten by telling them you have 3 dB more marbles than they have. Best regards Steffen Subject: Re: Prime factors of number near googolplexplex > Phil, > near googolplex ( http://www.alpertron.com.ar/GOOGOL.HTM ). This > thread is about the new page about factors of number near > googolplexplex ( http://www.alpertron.com.ar/GOOGOLP.HTM ). Of course, > prime factors of googolplexplex-1, -10 or +10 are very easy to find > using the methods you mention above. Aha, my window was too narrow!EItherthat ore I thought there was an echoecho. Just out of curiosity, I started generating factors of 10^10^10^100-1, and there aren't as many as I'd thought. The biggest so far is: 2900240811398724360969936242327094078063964843751 I'm generating them at about 5 seconds per prime, but that will slow down as they get larger. Shall I just mail you the whole log in a day, or do you want me to use the form on the web-page? Oooh: 27985968577282083202817375422455370426177978515625000001 To avoid anyone else duplicating my search space, it's: (10k+{1,3,7,9}) * 2^a * 5^b + 1 k=0..99; a+b<1000) 10^10^10^100-100 looks like a more interesting beast so I'll look at that next. However, 10^10^10^100-10 looks like it's a complete orgy of factors, and you may as well generate those locally, or your inbox will fill quicker than a quickly-filling thing. 10^10^10^100+100 looks interesting too. 10^10^10^100+1 looks, ahem, either interesting or incredibly boring. I won't be looking there. Ooooh: 410207670714926138814098853947048839430067346256691962480545043 94531251 Phil 393168275162152283817205516936940656869126087258337065577507019 04296875001 -- 1st bug in MS win2k source code found after 20 minutes: scanline.cpp 2nd and 3rd bug found after 10 more minutes: gethost.c Both non-exploitable. (The 2nd/3rd ones might be, depending on the CRTL) Subject: Re: ZZCrap In sci.math, ZZBunker : >> In sci.math, ZZBunker >> > : >>> In sci.math, ZZBunker >>> >> : >>> >> >>>>to isn't a verb. >>>> >>>> The last time I heard that to wasn't an intransitive >>>> verb was the last time that math wasn't spelled >>>> in the moronic fashion maths. >>> >>>> What's math? >>> >>> Nobody knows, Since like we've been telling morons >>> for several thousand years, now: >>> Maths is philosophy not science. >>> See Plato for philosophy and festive pillar work. >> >>> Math is an abstract model, nothing more. Occasionally it's >>> useful; we've even harnessed the Queen of Uselessness, >>> number theory, into modern cryptography. :-) Who says >>> primes aren't useful? :-) >> >>> However, '1' cannot be captured in a butterfly net. 'pi', >>> 'e', and i=sqrt(-1) are even more elusive. Good luck >>> finding a Taylor or MacLaurin series running around in the wild. >> >> Nobody ever said that you can find *anything* related >> to moronic *Calculus* running around in the wild. >> We have merely stated that you can find *real* >> *Turing machines* running around in the wild. > >> Models only, and imperfect ones at that. The ideal Turing >> machine has an infinite length tape (most of which is blank). >> All modern machines have finite tapes -- if they have tapes >> at all. > The ideal Turing machine never needed an infinite > tape to begin with. Since neither Turing nor any > other mathematician before him proved anything > conclusive about the word finite, other than > their token prayers to Euclid. True; running a Turing machine with an infinite empty tape would be slightly pointless, even were the machine the simple one: S=0: S <= 0, '0', Right S=0: S <= 1, '0', Right S=1: S <= 0, '1', Right However, from a theoretical standpoint one has to specify something along the lines of has a tape longer than needed. As for finite: everything's finite. Even the symbol '+oo' is a token of the concept we think of as infinity; it is not, however, infinite in itself. (The only infinite thing is God, if such exists at all, and even that's highly debatable -- but probably not here. :-) ) >> >> >>> There are examples of Nature taking advantage of Fibonacchi >>> numbers (e.g., sunflower seed spirals, plant stalk branchings) but >>> that doesn't mean Fibonacchi numbers are out there. >> >>> The main requirement for math is that it be self-consistent. >> >> The *only* requirement for math is that it >> exist in a non-existent vaccuum. >> Other than that it's the most trivial >> thing in the universe. > >> Math does not exist except as a series of thought patterns >> and as ink on paper and bits on storage devices and traveling >> through cables and air -- and those are also imperfect models. > Since models of math are also math, > we have never claimed that cables > are even approximately anything > other than exactly what they are, > which is cables. But when you > get somebody other than a > retarded mathematican using cables, you also > get a free subscription to > GPS TV, with the cable. So > again whatever math dorks > consider perfect is up to > math dorks. The rest of us are > still going to Mars. No > matter what Feynmann & heads Inc. say. Cables can be (at least) the following: [1] Information conduits. [2] Power conduits. [3] Garrotes. [4] Structural supports (e.g., suspension bridges) [5] Directional indicators (i.e., paths) Of course we are going to Mars. The question is when. (I'm not familiar with GPS TV or math dorks. Are these related to the talk.origins howler monkeys? :-) ) >> In short, there's no there there. It is merely a way of >> coloring one's perceptions of the Universe; without math, >> we'd look at the Universe differently -- perhaps using >> mysticism? -- and draw conclusions in a different fashion. > Well you've obviously never heard of Descartes. > He said the same thing and nobody believed him. > Since his use of the principle of extension, > was second only to Zeno's in being the > most overrated piece of philosophy ever written. That it is. I prefer, as Leonard Nimoy once said, the concrete, the graspable, the provable. I think, but amness (to coin a word) may not depend thereon, although on Usenet, one has to think to write, although the amount of thinking can be miniscule; I could try to be highly intelligent and string words into elegant sentence structure, or just exercise the minimum amount of thought necessary to hold down the 's' key and generate a string of sssssssssssssssssssssssssssssssssssssssssssssssssssssssssssss. .. until I decide to release the muscles holding down that key. (Most computers/computer keyboards have auto-repeat.) Infinity is not provable unless carefully specified. At best, it's mathematical sophistry for the statement for any integer N, there's always a bigger integer N+1. (which is a variant of weak induction). At worst, it's lazy thinking. There are admittedly some interesting questions, though; can one create/specify a Turing machine that, given an input tape and an alphabet {0,1,2,3,4,5,6,7,8,9,.,+,-}, with a valid number (encoded in the usual way), generate output with {0,1,2,3,4,5,6,7,8,9}, and for every unique input the output is valid (either '0' or begins with a nonzero) and unique? And is there another turing machine that reverses the process, and for every unique input the output is valid (begins with an optional sign, contains exactly one '.', and there's either a single '0', nothing, or a string of digits starting with nonzero before the '.') and unique? The existence of such machines would prove that the map between N and R would (a) exist, and (b) be computable. Most mathematicians would probably be of the opinion that such machines do not exist (see, for example, http://en.wikipedia.org/wiki/Cantor%27s_first_uncountability_ proof ) and might characterize the problem as the infinities are different. -- #191, ewill3@earthlink.net It's still legal to go .sigless. Subject: Re: SCHOENFELDS RANDOM THEOREM every possible finite sequence of digits occurs in pi (and most other > irrational numbers), many times, Does say 0320495307103004298010 qualify as a sequence of 8 'zeroes' or must the 00000000's be consecutive? regards bill j. Subject: Re: SCHOENFELDS RANDOM THEOREM <404fb4fd$0$8358$afc38c87@news.optusnet.com.au <7DO3c.71$2f4.61910@news.uswest.net >> Any finite sequence of digits of pi is truly random with a uniform >>> distribution. >>> Proof? URL? Anything? >>http://mathworld.wolfram.com/PiDigits.html >> Please quote the relevant part of that page. > Is your web browser broken? Is the copy-n-paste features of your browser/newsclient/window manager/whatever broken? You said that page proves your claim. Richard doesn't see how anything on that page does so[1]. Is it so terribly hard to point out which part of that text proves that Any finite sequence of digits of pi is truly random with a uniform distribution? Do you think you've proved the point? Footnotes: [1] Beats me if Richard is right or wrong. -- It's your choice though, if you do not believe in mathematics, in the importance of its healthiness and correctness, then you can just walk away now. -- James S Harris, on the Pythagorean Oath Subject: Re: SCHOENFELDS RANDOM THEOREM >>>> Any finite sequence of digits of pi is truly random with a uniform >>>> distribution. >>>> Proof? URL? Anything? >>>http://mathworld.wolfram.com/PiDigits.html >>> Please quote the relevant part of that page. >> Is your web browser broken? > Is the copy-n-paste features of your browser/newsclient/window > manager/whatever broken? > You said that page proves your claim. Richard doesn't see how > anything on that page does so[1]. Is it so terribly hard to point out > which part of that text proves that Any finite sequence of digits of > pi is truly random with a uniform distribution? > Do you think you've proved the point? What I see on the page is: It is not known if pi is normal (Wagon 1985, Bailey and Crandall 2001), Paraphrasing the question by asking if any finite sequence of digits of pi [is] truly random with a uniform distribution misses the point. Any finite sequence of digits of pi has a degenerate distribution with exactly one possibility. The first 5 digits are 31415 with probability 1. The 6th through 10th digits are 92653 with probability 1. The 11th through 15th are 58979 with probability 1. And so on. No distributions are involved unless you count these degenerate distributions. Using the term random seems to miss the point as well. It isn't relevant whether pi is random. What we're talking about is the limiting distribution of values of m digit contiguous sequences selected uniformly from the n-m possible starting points within the first n digits of pi as n increases without bound. This limiting distribution (if it exists) will be a discrete distribution. It may be uniform. Or it may be non-uniform. But it will not be random. John Briggs Subject: Re: SCHOENFELDS RANDOM THEOREM >Note that this is not a cop out as there should exist a way >>to prove this question (Unlike proving or disproving god). > Why should there exist a way to prove it? It might be undecidable. Because if it's undecidable, then this world is not the best of all possible worlds, which contradicts that God is omnipotent and all-good. Duh. -- Jesse Hughes Depression hits more people than thought. --headline in Lexington, KY newspaper, as reported on NPR's Morning Edition Subject: Re: SCHOENFELDS RANDOM THEOREM Originator: richard@cogsci.ed.ac.uk (Richard Tobin) >> Why should there exist a way to prove it? It might be undecidable. >Because if it's undecidable, then this world is not the best of all >possible worlds, which contradicts that God is omnipotent and >all-good. I'll apply for research funds from the Vatican immediately. -- Richard Subject: Re: SCHOENFELDS RANDOM THEOREM X-SessionID: fNa4c-6550-45-16359@news.uchicago.edu X-Hash-Info: post-filter,v:1.4 X-Hash: 8edfdc81 6d0270bf 6c788614 5ed82e5e 547bc3da >I'm posting because I'd hate to see a misunderstanding between two >decent people escalate into something uglier than it need be -- > >[big snip] >> There is nothing wrong with believing in what you're trying to prove >> as long as you don't confuse your belief for a proof. It is a fine >> line here, since in science >[snip comment by GR_Learner, as I want to focus on what Mati said] >> 1) Saying I believe this is true is a standard operating procedure. >> 2) Saying It is true since I believe in it marks you as a cracpot. > >> You obviously did not interpret what I said correctly. >And I think you misread Mati. It's quite clear he was not trying to >attack *you* personally, at least not here; his use of you here is >idiomatic English, meaning one. The context is very general; he's >trying to characterize *science*, not you. I though it is quite clear that this is the case. Mati Meron | When you argue with a fool, meron@cars.uchicago.edu | chances are he is doing just the same Subject: Re: SCHOENFELDS RANDOM THEOREM X-SessionID: fNa4c-6550-45-16359@news.uchicago.edu X-Hash-Info: post-filter,v:1.4 X-Hash: 0a19bb5a bcd07a8f 73384e51 fcf40a5b d1dc2a7d >> >>> >GR_Learner@GR.grv >>>> >>>>> >>>>> >>>>> >> Any finite sequence of digits of pi is truly random with >a >>>>uniform >>>>> >> distribution. >>>>> >>>>> >> Proof? URL? Anything? >>>>> >>>>> >http://mathworld.wolfram.com/PiDigits.html >>>>> >>>>> > Please quote the relevant part of that page. >>>>> >>>>> >>>>> >>>>> Quoting from the page GR_Learner@GR.grv referred you to >>>>> http://mathworld.wolfram.com/PiDigits.html >>>>> >>>>> The following distribution of decimal digits d is found for >the >>>first >>>significant >>>>> departure from a uniform distribution. >>>> >>>>> 10^n is not a statement about the entire series. >>>> >>>>What part of finite sequence of digits of pi don't you >understand? >>>>The first 10n is a finite sequence. >>>> >>>> There is *no* proof present in the above for statement made. A >>>> specific sequence is *not* any specific sequence. End of story. >>> >>>So... do you think that the first 10^n decimals are special in that >they >>>show no significant departure from a uniform distribution? >> >>> The answer is, you don't know. >> >>That is true of all things. However, Occam's razor will allow me to >place a >>bet with confidence. If Kip can bet a bottle of wine on Black Holes, I >>could do the same for this. >> You can place bets with whatever confidence you wish but a bet is not >> a proof either. >Prove to me it is not. :-))) >>> The sequence a_n = exp(kn)/n decreases for n < (1/k), then it starts >>> increasing and tends to infinity. picking k of, say, 10^(-1000), >>> you'll be going a very long way seeing it decreasing, before it starts >>> increasing. After the first 10 billion or so terms, you'll sure be >>> ready to state that, this one tends to zero. But, it ain't. >> >>> A finite number of cases does not make for a proof, in math. It may >>> inspire a quest for a proof, but that's all. >> >>This is a statment of the statistical measruments of the randomness of >Pi. >>http://www.super-computing.org/pi-decimal_current.html >> >> What part of the word proof you don't understand? >Let me ask you somthing... do you apply the same standard or definition of >proof in physics as you do in math? No. You don't prove things in physics, you only confirm the validity of models within a given range of paramters and to given tolerances. > Does either have more merit in describing reality? Math, per se, does not describe reality. it is a set of logical models. You can apply mappings between some of these models and some observable aspects of reality but then the correctness of the mappings is a separate issue from the correctness of the math theorems. >Note that in physics, a proof of the validity of a model as it applies to >reality can never be found. Indeed. >Even if there is a absolute proof in certain aspects of mathematics (and >this is not the case for all aspects as you know) that does not mean it has >any validity for reality. Just what I said. >As a physicist, there is a certain point where with enough data I would say >that the theory is proved You were doing well so far, now you blew it again. No, there is no point where a physical theory is proved, nor can there be such point. Mind you, Newtonian mechanics passed with flying colors all possible tests for upward of 200 years. Until, the point has been reached that it didn't. >... make use of the results... and move on (Until >there is evidence that the theory does not apply). Meaning, it wasn't proved. I'm afraid you just don't understand what proved means. > The data is beyond convincing in the case of pi. Two points: 1) There are known cases of mathematical properties which only become true for truly astronomic numbers. So your convincing is meaningless. 2) More important, we're not debating here , one opinion against another. There are acknowledged standards of what is and what isn't a mathematical proof. You're not being asked for your opinion of what they should be. You're just being told what they're. Now, you made a specific statement and provided as a proof something which is anything but. I suggest you acknowledge this and move on. I've neither the time, nor the inclination for Dionysian debates. Mati Meron | When you argue with a fool, meron@cars.uchicago.edu | chances are he is doing just the same Subject: Re: SCHOENFELDS RANDOM THEOREM X-SessionID: fNa4c-6550-45-16359@news.uchicago.edu X-Hash-Info: post-filter,v:1.4 X-Hash: f5eca798 dfc04500 7856c7db 9298e031 6ebca380 >> The answer is, you don't know. > >> The sequence a_n = exp(kn)/n decreases for n < (1/k), then it starts >> increasing and tends to infinity. picking k of, say, 10^(-1000), >> you'll be going a very long way seeing it decreasing, before it starts >> increasing. After the first 10 billion or so terms, you'll sure be >> ready to state that, this one tends to zero. But, it ain't. > >> A finite number of cases does not make for a proof, in math. It may >> inspire a quest for a proof, but that's all. >All of which brings to mind an old joke: >Physicist's induction: 3 is prime, 5 is prime, 7 is prime; therefore all >odd numbers are prime >Engineer's induction: 3 is prime; therefore all odd numbers are prime >Chemist's induction: 3 is prime, 5 is prime, 7 is prime, 9 is prime, 11 is >prime; therefore all odd numbers are prime :-))) And oldone but still a good one. I recall that there was a variation including a computer scientist, as well, but I forgot the details. Mati Meron | When you argue with a fool, meron@cars.uchicago.edu | chances are he is doing just the same Subject: Re: SCHOENFELDS RANDOM THEOREM X-SessionID: fNa4c-6550-45-16359@news.uchicago.edu X-Hash-Info: post-filter,v:1.4 X-Hash: bee46d54 1b29faca a729adfd 9585abe7 75bfbbbf >[...] >> It may strongly suggest so, but this still is not a proof. >>You might say show me the proof and I can with equal validity say show me >>the disproof. >He made the affirmative statement, ________ is... Burden of proof is his. >Your only affirmative statement was that he had not proven his. You met your >burden of proof by demonstrating his data was not logical evidence for his >unqualified claim. Yes, for sure. I think that he may simply not be aware that math uses a precise language in which .. it is ... means much more than just I think it is... or I believe it is.... >> Sorry, that's not how math works. >Not how the logic of verbal statements works, either. >Sorry about horning in on your thread, but the poster seems to be avoiding >my replies. I wonder why. :-) I wonder too:-) Still, I've seen many cases in which the order in which posts appear on one server differ (quite significantly, at times) from they time ordering on another. So, give it some more time. Mati Meron | When you argue with a fool, meron@cars.uchicago.edu | chances are he is doing just the same Subject: Re: SCHOENFELDS RANDOM THEOREM X-SessionID: Nso4c-8668-45-22794@news.uchicago.edu X-Hash-Info: post-filter,v:1.4 X-Hash: f198cb1e e31ccbe6 90d0a26b 0ace8a76 00b4d845 >> >>[...] >> >>> It may strongly suggest so, but this still is not a proof. >> >>>You might say show me the proof and I can with equal validity say >show me >>>the disproof. >> >>He made the affirmative statement, ________ is... Burden of proof is >his. >> >>Your only affirmative statement was that he had not proven his. You met >your >>burden of proof by demonstrating his data was not logical evidence for >his >>unqualified claim. >> >> Yes, for sure. I think that he may simply not be aware that math uses >> a precise language in which .. it is ... means much more than just >I think it is... or I believe it is.... >> >>> Sorry, that's not how math works. >> >>Not how the logic of verbal statements works, either. >> >>Sorry about horning in on your thread, but the poster seems to be >avoiding >>my replies. I wonder why. :-) >> >> I wonder too:-) Still, I've seen many cases in which the order in >> which posts appear on one server differ (quite significantly, at >> times) from they time ordering on another. So, give it some more >> time. >I killfiled Servo if that is what you are wondering. There is no accounting for taste. Mati Meron | When you argue with a fool, meron@cars.uchicago.edu | chances are he is doing just the same Subject: Re: SCHOENFELDS RANDOM THEOREM [...] >>Sorry about horning in on your thread, but the poster seems to be >>avoiding my replies. I wonder why. :-) >> >> I wonder too:-) Still, I've seen many cases in which the order in >> which posts appear on one server differ (quite significantly, at >> times) from they time ordering on another. So, give it some more >> time. >I killfiled Servo if that is what you are wondering. > There is no accounting for taste. True. GR_Learner is the one who responded to Evil Al's 100 lines ROTFLMAO! Imagine the fun he'll have when /Mein Kampf/ is published in comic book form. Being killfiled by someone who finds the diseased sewage of Carbuncle Al funny is an honor. It saddens me only insofar as there is a slight chance of redemption for GR_Learner given sufficient education, against which he innoculates himself by excommunicating the opposition. (That last sentence reminds me of a problem I have with usenet and gender-specific pronouns. GR_Learner could be a female, for all I know. And, so could Mati. Or Servo. If so, my apologies to anyone who may be sensitive to being mislabeled while still preferring gender anonymity. English can be most annoying in its deficiencies. It is not widely regarded as a polite substitute. I deal with this regularly in my writing, but will be damned before I rethink every sentence of usenet postings. Too much like work.) Servo Subject: Re: SCHOENFELDS RANDOM THEOREM >[...] >> It may strongly suggest so, but this still is not a proof. > >>You might say show me the proof and I can with equal validity say show me >>the disproof. >He made the affirmative statement, ________ is... Burden of proof is his. >Your only affirmative statement was that he had not proven his. You met your >burden of proof by demonstrating his data was not logical evidence for his >unqualified claim. > Yes, for sure. I think that he may simply not be aware that math uses > a precise language in which .. it is ... means much more than just I think it is... or I believe it is.... >> Sorry, that's not how math works. >Not how the logic of verbal statements works, either. >Sorry about horning in on your thread, but the poster seems to be avoiding >my replies. I wonder why. :-) > I wonder too:-) Still, I've seen many cases in which the order in > which posts appear on one server differ (quite significantly, at > times) from they time ordering on another. So, give it some more > time. I killfiled Servo if that is what you are wondering. Subject: Re: SCHOENFELDS RANDOM THEOREM [...] > I wonder too:-) Still, I've seen many cases in which the order in > which posts appear on one server differ (quite significantly, at > times) from they time ordering on another. So, give it some more > time. > I killfiled Servo if that is what you are wondering. Oh, THAT'S right, this was the little fellow who was lecturing me on just *PLONK*ing Evil Al if I don't like the bile he spews. His loss. Servo Subject: Re: SCHOENFELDS RANDOM THEOREM [...] > You can place bets with whatever confidence you wish but a bet is not > a proof either. > Prove to me it is not. You're becoming incoherent. Do you wish Mati to prove that a bet is not a proof, of that some series of pi is not truly random? [...] >This is a statment of the statistical measruments of the randomness of >Pi. >http://www.super-computing.org/pi-decimal_current.html > > What part of the word proof you don't understand? > Let me ask you somthing... do you apply the same standard or definition of > proof in physics as you do in math? No. > Does either have more merit in describing reality? Mathematics is not concerned with physical realities. Mathematics is a broad- ranging field of study in which the properties and interactions of idealized objects are examined. http://mathworld.wolfram.com/Mathematics.html > Even if there is a absolute proof in certain aspects of mathematics (and > this is not the case for all aspects as you know) that does not mean it has > any validity for reality. It does for mathematical reality. [Snip rehash of empiricsm...] > The data is beyond convincing in the case of pi. Different folks and applications have different requirements for what might be convincing. In US criminal law it's reasonable doubt, while in civil law it's preponderance. You may be convinced by the preponderance of evidence, but you are still offering an UNPROVEN assertion if you say Any finite sequence of digits of pi is truly random with a uniform distribution. You might decide from a preponderance of evidence in civil court that eight accountants are guilty of embezzlement, but would be committing the fallacy of unwarranted generalization if you said that 'Any accountant is a crook.' Physics is based upon a preponderance of evidence. Mathematics draws conclusions which are usually beyond a reasonable doubt. > Now a bit of a digression to ask you a question... feel free to bring back > on topic if you wish. > I am not certain that there is any proof of randomness for any finite > sequence of numbers. Is this true? > Also, I am not certain if there is any test that can be done to determin if > a finite sequence of number generated by any mathematical procedure or > algorythem is random. We've been beating this to death in other threads. Definition matters. > In fact, I have read that true randomness is can only be properly defined > and described by the actiions of natural systems, such as radioactive decay, > and that there is not mathematical procedure to generate a truley random > sequence. Google on chaos theory, BBP algorith and the pi series. That nature thing is no foregone conclusion, either. > Now, if this is true, then perhaps the definition of proof for the > randomness of pi's digits should be of the emperical type, not the > mathematical? No problem, if you say Empirical evidence suggests that 'any finite sequence of digits of pi is truly random with a uniform distribution,' with the clear understanding that any counterexample would falsify that belief. That's empirical science. Still does not make the unqualified Any finite sequence of digits of pi is truly random with a uniform distribution a legitimate assertion, however. > You obviously did not interpret what I said correctly. The fault for > that is my fault. No, he interpreted the word any with complete accuracy. True, though, saying any was your fault. [Snip personal silliness.] > Fermat's last theory. Wiles would not have bothered if he did not believe > that it was true. In a colloquium I had the privilege to sit in on he > stated that since he was a kid he believed that FLT was true... that he knew in his heart of hearts that the theory could not possibly been > incorrect, and that this belief propelled him into a career in mathematics > and the ultimate solution of the problem. All of which says nothing about the truth of FLT or the randomness of the pi series. The history of mathematics is replete with the corpses of those who were wrong about their hunches. Even astrologers and stopped clocks are right once in a while. > You mean that to disprove a theory is not valid in mathematics? > Mati... please prove to me that any finite sequence of the decimals of pi is > NOT truly random. Fallacy of shifting the burden of proof. He NEVER said it wasn't--only that your assertion of any was unproven. > In fact, I will be satisfied if you could outline how such a proof would be > attempted. Hell, so would Mati. A Fields Medal nomination would surely be in order. Following your line of thinking, I could assert that there are an infinite number of twin primes. After all, they're everywhere we look so far. I could demand, when challenged, that my critic produce a proof that there AREN'T. I would be: 1) making a logically unproven statement, and 2) diverting attention from my error using the fallacy of shifting the burden of proof. > So... do you think I am a fool? One sure way to prove you're not would be to admit that your unqualified use of any was an error or misstatement. It's not hard--I do it all the time. Two things will each lower our IQ thirty points over a lifetime--avoiding admission of errors, and raising children. Servo Subject: Re: SCHOENFELDS RANDOM THEOREM > Trivially disproven by example in Hofstadter's Godel, Escher, Bach. > It's only 777 pages long. If you look at each page for one second you > can find the table within 12 minutes. > Uncle Al is wrong, and transistively, so are you. Try independent thought. > No, Uncle Al is right, as he very often (but not always) is, and you > are wrong. All you can prove is that the probability of your statement > being correct is 1, and conversely, the probability of it being false > is 0. > Now, if x is a number chosen at random out of the set of all > intergers/rational numbers/real numbers, the probability that x=0 is > zero. Have I proved that 0 does not exist? I said that R the random sequence is an INTEGER sequence BOUND by n and m. Let R = infinite sequence of random integers bound by n and m Let x = an integer such that n <= x <= m The probability of drawing x from R at any point is 1/(m-n + 1) The probability of NOT drawing x from r at any point is 1 - 1/(m-n + 1) Let x be the last element of some arbitrary sequence of integers S such that for all y in S, n <= y <= m The probability of drawing S from R starting at any point is 1/(m-n + 1)^|S| The probability of NOT drawing S from R starting at any point is 1 - 1/(m-n + 1)^|S| Now, the probability of NEVER drawing S from R is lim j->|R| (1 - 1/(m-n + 1)^|S|)^j Now let T = 1/(m-n + 1)^|S| We see that 0 < T < 1 We also see that 0 < 1 - T < 1 Since |R| = +inf, Then, lim j->+inf (1 - T)^j = 0 Therefore the probability that S does not occur in R is 0, transistively, the probability that S DOES occur in R is 1 - 0 = 1. PROOF > I'll drink to transistivity (hic!), especially of the bipolar junction > kind, > Cheers, > Zigoteau. JS Subject: Re: SCHOENFELDS RANDOM THEOREM > Therefore the probability that S does not occur in R is 0, > transistively, the probability that S DOES occur in R is 1 - 0 = 1. > PROOF With the greatest of respect, that is exactly what I thought your proof would end up demonstrating, and in fact by your use of the word 'random', it was the only sort of thing it _could_ end up demonstrating. However if you choose an element at random from an infinite set, the probability of choosing any particular element is zero. Hence proving that an event has probability zero does not prove that it is impossible. Looks like the Nobel Prize for Mathematics has to be postponed for yet another year. However, transistively speaking, you might get the Nobel Prize for Electronics. Cheers, Zigoteau. Subject: Re: SCHOENFELDS RANDOM THEOREM I'm posting because I'd hate to see a misunderstanding between two decent people escalate into something uglier than it need be -- [big snip] > There is nothing wrong with believing in what you're trying to prove > as long as you don't confuse your belief for a proof. It is a fine > line here, since in science [snip comment by GR_Learner, as I want to focus on what Mati said] > 1) Saying I believe this is true is a standard operating procedure. > 2) Saying It is true since I believe in it marks you as a cracpot. > You obviously did not interpret what I said correctly. And I think you misread Mati. It's quite clear he was not trying to attack *you* personally, at least not here; his use of you here is idiomatic English, meaning one. The context is very general; he's trying to characterize *science*, not you. > It is much worse than a cop out. Mathematiclly speaking, that's ... > well, I don't want to get offensive, yet. > You already have become ofensive Mati. No, that was your misreading. I will look past that. Also... > please indicate where you have snipped text so it will not be taken out of > context. > Mati Meron | When you argue with a fool, > meron@cars.uchicago.edu | chances are he is doing just the same > So... do you think I am a fool? It's his *.sig*, for goshsakes. It appears in all of Mati's postings. Don't take everything so personally. (OTOH the things you are saying here about mathematical proof *are* foolish. Asking Mati to Prove... that a bet isn't a proof? That's just silly. Smart people say foolish things sometimes, and I would say that for you this is one of those times. IMO.) Subject: Re: SCHOENFELDS RANDOM THEOREM X-SessionID: fNa4c-6550-45-16359@news.uchicago.edu X-Hash-Info: post-filter,v:1.4 X-Hash: 0746b2e5 33b1ebb2 34205f41 26a4b484 6a4cb538 >> I'm posting because I'd hate to see a misunderstanding between two >> decent people escalate into something uglier than it need be -- >Thanks. I stand corrected with regard to saying that the digits of pi are >randomly distributed. Matis is right that no matter how many digits are >sampled that still does not constitute a mathematical proof. >My apologies if I insulted Mati. You didn't. I was trying to set the terminology straight since this issue, of proof versus confirmation versus indication etc. is one of the eternal sources of confusion on this ng. So, I'm glad we got it cleared up. Mati Meron | When you argue with a fool, meron@cars.uchicago.edu | chances are he is doing just the same Subject: Re: SCHOENFELDS RANDOM THEOREM > I'm posting because I'd hate to see a misunderstanding between two > decent people escalate into something uglier than it need be -- Thanks. I stand corrected with regard to saying that the digits of pi are randomly distributed. Matis is right that no matter how many digits are sampled that still does not constitute a mathematical proof. My apologies if I insulted Mati. Subject: Re: SCHOENFELDS RANDOM THEOREM > I'm posting because I'd hate to see a misunderstanding between two > decent people escalate into something uglier than it need be -- > Thanks. I stand corrected with regard to saying that the digits of pi are > randomly distributed. Matis is right that no matter how many digits are > sampled that still does not constitute a mathematical proof. > My apologies if I insulted Mati. Well done. More should follow this example. Servo Subject: expectation of the product of two dependent random variables Hi sorry to bother you guys here. Let random variable X = a_1r_1+a_2r_2+...+a_nr_n, and random variable Y = b_1r_1+b_2r_2+...+b_nr_n, where a_i and b_i (i=1...n) are constants, and r_i is i.i.d. random variable chosen from N(0,1). Note that r_i in both X and Y are the same. So how to compute the expectation of exp(p times X times Y), that is E[exp(p times X times Y)] = ? Thanks a lot roy Subject: re:expectation of the product of two dependent random variables > exp(p times X times Y) Please clarify what this expression means. Posted Via Usenet.com Premium Usenet Newsgroup Services ---------------------------------------------------------- ** SPEED ** RETENTION ** COMPLETION ** ANONYMITY ** ---------------------------------------------------------- http://www.usenet.com Subject: Re: (-2/3)^(-2/3) = (3/2)^(2/3)? > Note that GrafEq plots no points for y = (-2)^(x^x) where -1 < x < 0. > Is there a mathematical difference between the systems: > |y = (-2)^u |y = (-2)^(x^x) > |u = x^x and |-1 < x < 0 > |-1 < x < 0 > According to GrafEq the graph of the first system consists of four > curves and infinite vertical line (probably x = 0). The graph of the > second consists of point (probably (-1;-0.5)) and infinite vertical > line (probably x = 0). > So everything is OK? > As far as I can tell, everything is OK with GrafEq. But your statements > about what the graphs consist of according to GrafEq are subtly incorrect. I agree with the last sentance. But there are cases in which GrafEq doesn't plot correctly (it doesn't plot at all) the graph of mathematical equation or system. For example from the following system: |u * u = 2 |x = ((-3)^u)^u |x belongs to R => x = -9. Right? On that basis the graph of y = ((-2)^sqrt(2))^sqrt(x) should be visually identical to y = -(2^sqrt(2))^sqrt(x). Right? Calvin Subject: re:cantor's theorem I dont agree yet to the old hat: As far as I know uses different types of sets, all belonging to different levels. In my set theory, there is only one kind or type of set. But I use types or levels for the element relation. In my theory, a set can be element of itself - ins theory that is not allowed. But my meta set theory is not completely ready yet, as for example I could construct the natural numbers as sets, but not the set of all natural numbers... Posted Via Usenet.com Premium Usenet Newsgroup Services ---------------------------------------------------------- ** SPEED ** RETENTION ** COMPLETION ** ANONYMITY ** ---------------------------------------------------------- http://www.usenet.com Subject: Re: puzzle: GCDs of Infinite Set of Integer Pairs >>)>If you pick two numbers randomly from [0..1], is it possible that those >>)>two numbers are the same ? >>) Yes. But it is not possible that either of those numbers is 1.5. >>So even though they won't be the same no matter how many times you try, >>you say that it's still possible ? Hey I went away for 3 weeks to NZ, and you guys have managed to turn this into one of the longest threads in the history of usenet, but AFAICS you still haven't resolved the basic point for me, even though the discussion has got back to pretty well exactly the same place. My contention is: You can say 'select a real randomly from the range [0,1]' but you can't actually do it. It's like saying 'Imagine you can travel faster than the speed of light'. A fine thought experiment, and you can draw all sorts of conclusions about a universe that allows it, but this is not that universe. I contend that: 1. There is no physical way of doing it (spin a pencil etc fails due to the quantum nature of reality) 2. There is no way to describe a finite process which results in selecting such a number randomly 3. There is no way to accurately describe the number you selected. If you describe the number precisely, I'm entitled to call you a liar, and assert with certainty that you didn't choose randomly. On this basis, I assert that it's overly optimistic at best to say choose a real number at random in the interval [0,1] Oh, and I can't agree with this obviously: >Who says they won't be the same no matter how many times I try? That's >the way to bet, but it's not a certainty. If ever there was a certainty, this is it. You can't even choose a number *once* let alone twice. Even if I am eventually persuaded that you can in fact choose a number at all, you definitely can't pick it again. 'Probability zero' has to count for something in our little mathematical universe, or it's a useless little universe. -- Patrick Hamlyn posting from Perth, Western Australia Windsurfing capital of the Southern Hemisphere Moderator: polyforms group (polyforms-subscribe@egroups.com) Subject: Re: General algorithm for vector cross-product by support1.mathforum.org (8.11.6/8.11.6/The Math Forum, $Revision: 1.9 primary) id i2C1JVE04033; The cross product is a convenient way to find a vector that is normal to the plane spanned by two non-collinear vectors, in 3 dimensions. Just as Don mentioned, IF you have n-1, n dimensional vectors, then the vector that is perpendicular to these is found by computing the determinant of the n x n matrix consisting of the n-1 vectors and the e_1, e_2, ..., e_n basis. However, say I have two 'k' dimensional vectors and I want to find the vector that is perpendicular to them. In general, there is no _one_ unique vector. In fact, in general, this problem usually has a solution that is a vector space. In particular, if I have m non-collinear 'n' domensional vectors (with mHey all, >Alas, there doesn't seem to be a good newsgroup for mathematical algorithms, >so I'm posting here. >I am using the cross-product to find a normal that us perpendicular to two >vectors. All of the examples and code snippets that I have ween deal with >the 3D case. Correct me if I am wrong, but we can find the cross-product of >other dimensions (either lower or higher), can we not? >To handle higher dimensions, we could recursively find the determinent of >the vectors, but if we have n-dimensional vectors, do we need n-1 vectors to >fill the n x n matrix? >Also, how would we lower dimensions? I imagine that the cross of >one-dimensional vectors (i.e. numbers) is just their arithmetic product. >What about R^2? >So is there some general algorithm for handling all cases? >Dave Subject: Intergrals by support1.mathforum.org (8.11.6/8.11.6/The Math Forum, $Revision: 1.9 primary) id i2C1JVr04029; int [sqrt(36-x^2)] dx help! i dont know wat to substitute x as! Subject: Re: Intergrals by support1.mathforum.org (8.11.6/8.11.6/The Math Forum, $Revision: 1.9 primary) id i2CH8lQ06002; >int [sqrt(36-x^2)] dx >help! i dont know wat to substitute x as! You mean without actually looking in a text book? :) Try the substitution 6x= sin(theta). The point of that is then sqrt(36-x^2)= sqrt(36(1-sin^2(theta))= 6cos(theta). Also 6 dx = cos(theta)dtheta or dx= (1/6)cos(theta)dtheta so the integral becomes int cos^2(theta)dtheta. Use the trig identity cos^2(theta)= (1/2)(1+ cos(2theta)) to do that integral. Any calculus text book should have a section on trigonometric substitutions in the chapter of methods of integration. Subject: Re: Intergrals >int [sqrt(36-x^2)] dx >help! i dont know wat to substitute x as! > You mean without actually looking in a text book? :) > Try the substitution 6x= sin(theta). That would be counterproductive. Try x = 6*sin(theta). > The point of that is then > sqrt(36-x^2)= sqrt(36(1-sin^2(theta))= 6cos(theta). > Also 6 dx = cos(theta)dtheta or dx= (1/6)cos(theta)dtheta so the > integral becomes int cos^2(theta)dtheta. or indeed 6 times this. > Use the trig identity > cos^2(theta)= (1/2)(1+ cos(2theta)) to do that integral. > Any calculus text book should have a section on trigonometric > substitutions in the chapter of methods of integration. -- P.A.C. Smith The vast majority of Iraqis want to live in a peaceful, free world. And we will find these people and we will bring them to justice. Subject: Re: Intergrals > int [sqrt(36-x^2)] dx > help! i dont know wat to substitute x as! Try this: let 6 (sqrt(36)) be the hypotenuse and x the opposite side. It's not guaranteed to work (well, not all the time, with all possible combinations of stuff under the square root sign), but you can keep messing around with it and eventually something will work. Jon Miller Subject: Re: Intergrals Content-transfer-encoding: 8bit > int [sqrt(36-x^2)] dx > help! i dont know wat to substitute x as! x = 6sin(v) -- Paul Sperry Columbia, SC (USA) Subject: Re: sequence of rationals > Let {x_n} be a sequence of rational numbers that converges to an > irrational number, x. i.e. x_n -> x. Then the denominators of each > of the x_n must necessarily grow as n goes to infinity. > This fact is obviously true, but I am having a hard time proving it. > Thanks for any help. Assume that M is an integer upperbound on the denominators. Then M!*x_n, where M! is factorial M, must be an integer for each n, so we have the sequence of integers {M!*x_n} converging to the still irrational value M!*x. I don't think so. Subject: need help too (probability) Let X1, Z be independent where X1 is N(0; 1) and P(Z = 1) = P(Z = ?1) = 1/2. Let X2 = X1Z. Show that X2 is N(0; 1), and that X1 and X2 are uncorrelated (that is, have zero covariance) but not independent. Why does this not contradict the result given in class, that uncorrelated multivariate normal random variables are independent? Is there any clue to show that X2 is also N(0,1) How come X1 and X2 are uncorrelated but not independent? By the definition if the covariance is zero, correlation should also be zero. And they are independent. But why it is not? Can anyone help me ~~???? Thanks! Subject: Re: Non-AP Calculus? > Should high schools teach Regular Calculus courses for students who > are not at the level to handle the Advanced Placement Calculus course? > My instinct is to say that those students would benefit more from > additional precalculus preparation. Is that the way college > professors would prefer things, or would they rather have students who > have seen some calculus in a watered-down class. > I ask this question because I am the mathematics department chair at > my school, and we are looking at what courses we teach and what > courses we should teach. > Thank you for any insights you can provide! > Michael Brown I think it is a good idea. As a matter of fact, I did it for about 25 years. The non-AP course that we taught covered the AB curriculum plus many of the BC topics. We covered through advanced methods of integration, applications of definite integrals including arc length, surface area, parametric equations. Basically what was skipped were infinite series, power series, and ODE's that used to be a bigger part of the BC topics list. The difference was that we did not have to be done by the time that the AP exam was offered. The additional time allowed for better development and reinforcement. The students who took this course did one of three things when entering college: a) Skipped one semester, b) Entered an honors or more rapidly paced program, or c) Did the whole course over. Students in all of the above groups reported back positively about the experience. One student who had struggled to get a C- took his calc at a prestige school as a pass/fail because he was worried about his HS course. He got an A+ and could only call it a pass. The confidence developed by most seems to have made a big difference and of course seeing things for a second time helped. Also, the additional contact hours available in a HS setting allows time for firming up ideas from the pre-calc topics. This done in context of a calc environment seemed to work better for us than simply redoing pre-calc topics alone. It worked extremely well for me/us and the students. Ken Subject: Re: Non-AP Calculus? > Should high schools teach Regular Calculus courses for students who > are not at the level to handle the Advanced Placement Calculus course? > My instinct is to say that those students would benefit more from > additional precalculus preparation. Is that the way college > professors would prefer things, or would they rather have students who > have seen some calculus in a watered-down class. Why? Seriously, the AB course covers the first semester of college calculus in a year. That means that a student who is prepared can go to class and do an average of about half an hour of homework a day (more during the push for the exam). If they aren't willing to work that hard, will they be prepared for the class? So will they get anything out of the watered-down version? I think the answer to both those questions is no. If they aren't ready for calculus, they should either do more to get ready for calculus or study something that they find more interesting. Jon Miller Subject: Re: Non-AP Calculus? Content-transfer-encoding: 8bit > Should high schools teach Regular Calculus courses for students who > are not at the level to handle the Advanced Placement Calculus course? > My instinct is to say that those students would benefit more from > additional precalculus preparation. Is that the way college > professors would prefer things, or would they rather have students who > have seen some calculus in a watered-down class. > I ask this question because I am the mathematics department chair at > my school, and we are looking at what courses we teach and what > courses we should teach. > Thank you for any insights you can provide! > Michael Brown I can't claim to speak for all College/University Calculus instructors, but I think that many would prefer _no_ High School Calculus - AP or otherwise. While there are certainly exceptions, it is the a little knowledge is a dangerous thing syndrome. Students coming out of AP Calculus tend to think they know more than they actually do. The AP students, having coasted through AP Calculus, think that they will continue coasting in College/University Calculus and are sometimes slow to wake up. I think a solid trig course with _proofs_, a solid geometry course with _proofs_ and a very solid pre-calc algebra course would serve your students much better. You might want to ask your question on the sci.math newsgroup. -- Paul Sperry Columbia, SC (USA) Subject: Re: Non-AP Calculus? > Should high schools teach Regular Calculus courses for students who > are not at the level to handle the Advanced Placement Calculus course? > My instinct is to say that those students would benefit more from > additional precalculus preparation. Is that the way college > professors would prefer things, or would they rather have students who > have seen some calculus in a watered-down class. What would you consider to be appropriate precalculus preparation? Subject: Re: generalization of stokes' and divergence theorems? by support1.mathforum.org (8.11.6/8.11.6/The Math Forum, $Revision: 1.9 primary) id i2CH8kd05998; >Stokes theorem says SS[ curl F dS ] over S = S[ F dr ] over C, and the >divergence theorem says SSS[ div F dV] over E = SS[ F dS ] over S. >Is there a theorem that says, SSSS[ something(F) dH ] over H = SSS[ F dV] >over V. Wouldn't V have to have an orientation? It seems that in R^3 we >couldn't have that, but in R^4 I'd think you could define something like >that. >Is there a general something where I can write: >SS...n[ something(F) dA ] over A = SS...(n-1)[ F dB] over B >or something similar? >Thanks, >Jeremy The most general something is Poincare's formulation: Let M be a n-dimensional manifold with boundary dM (so dM is an (n-1)-dimensional manifold. Let w be an (n-1)-order differential form on dM with derivative (co-boundary) dw (an n-order differential form). If M is simply connected then the integral of dw over M is equal to the integral of w over M. If M is not simply connected, then the difference between the two integrals is related to the homotopy groups. Subject: Re: tensors by support1.mathforum.org (8.11.6/8.11.6/The Math Forum, $Revision: 1.9 primary) id i2CHagW09393; >I'm understanding tensors are things who obey certain transformation >properties and algebraic properties and everything, but I don't understand >what they physically/geometrically mean. What is the physical significance >of a covariant, contravariant, or mixed tensor of a specific order? I.e., >what's the physical difference between A^ij, A_ij, and A^i_j? I'm also >confused as to why we write a dot product of vectors in tensor notation as >a_i b^i, why not a_i b_i, or a^i b^i? Is there a difference between a^i b_i >and a_i b^i? I guess I'm not really understanding the difference between >covariant, contravariant, and mixed, other than that the differences in how >they transform. >I'd really appreciate it if someone could help me here! >Thanks, >Jeremy That's asking a lot! The problem is that you are used to working with vectors in Euclidean space with a cartesian coordinate system. Tensors are regularly applied to spaces or surfaces in which there is NO cartesian coordinate system (for example, the surface of a sphere). If you were to work only in Euclidean space and always used cartesian coordinate systems, there would be no difference at all between the contravariant and covariant coordinates of a tensor. Notice I said contravariant and covariant coordinates, NOT contravariant and covariant tensors. Strictly speaking any tensor has both covariant and contravariant coordinates. Here's a simple example. Take flat, two dimensional, Euclidean space and take the x-axis as the u coordinate axis but take the line y= x as the v coordinate axis. Now there are two different ways of labelling any point (u,v): (1), drop a perpendicular from the point to each axis and take the distance to the v-axis as the u coordinate and take the distance to the u-axis as the v coordinate. (2) draw lines through the point parallel to the coordinate axes. Measure the distance along the line parallel to the u-axis to the v-axis and make that the u coordinate and similarly for the v coordinate. In cartesian coordinates, of course, those would give exactly the same thing. In skew coordinates they are not the same and are the contravariant and covariant coordinates respectively. As for u_i v^i as dot product, do you know what the dual space is? For any vector space, the dual space is the set of all linear functions from the vector space to the real numbers. That is, if v is a vector and f a dual vector, then f(v) is a number. One can show that, in Euclidean space (i.e. flat with cartesian coordinates), the dual space is isomorphic to the vector space. That is, there is a natural (given the coordinate system) way to associate each vector with such a function. When we right u.v we really mean f(v) where f is the function associated that way with u. In more general spaces, there is no such association, but we CAN associate the contravariant components with functions on the covariant components and vice-versa. Subject: toplogy by support1.mathforum.org (8.11.6/8.11.6/The Math Forum, $Revision: 1.9 primary) id i2CJ6Y119302; Hello all Just like to know something about toplogy:- 1) Where i can find information about an Old magic which magician use to do in which they wear e.g. two shirts .... one is black ( or of any colour) and over that they wear a green shirt ( or of any color this is just an example) and by a topological trick they were able to takeoff the inner shirt in our case the black one... how this magic trick is related to toplogy? as iam using internet cafe ..not mailing from my own pc so i request u all if u have the link or info plz mail me... thankyou waqar Subject: Computer Science student seeks mathematic aid by support1.mathforum.org (8.11.6/8.11.6/The Math Forum, $Revision: 1.9 primary) id i2CJ6Yu19292; Since I dont have a background in math, I figure that I can more easily describe my question with a little practical background. I'm creating a program to find curves; three dimensional curves with specific endpoints and slope at those endpoints. The flow of the program will be this: Input. 1. Endpoints 2. Three dimensional unit vector at each endpoint describing the motion of the curve at that endpoint. 3. Length of the curve. Output. 1. Curve fuction (My biggest problem. I dont know if it is possible to create one function or colleciton of functions from the limited input. I've mildly studied polynomial interpolation and splines) Requirements. 1. Goes through both endpoints 2. Path of the curve must be along the unit vector at the endpoint 3. Must have stated curve length. The optimal curve will : 1. Be smooth 2. Minimize concavity 3. Minimize rate of change of concavity (I'm just saying the less curve the better. But the parts that match up to the unit vector at the endpoints need to be smooth. More broad curves are better than fewer sharper curves, too.) Can anyone point me in any direction? I'm currently in design time with the program and I need to find out my options. I dont mind doing guess-n-check for these values either. So if someone can give me a template for a curve which fulfills the requirements, I can use trial and error to find the most optimal of these curves. I didn't know where to look for help other than here. I hope you all find enjoyment on helping me with this project. I've talked to people at my university but those were student tutors, not professors. Thanks all in advance! Derek Adams Student, Computer Science Cal Poly Pomona Subject: Re: Computer Science student seeks mathematic aid > Since I dont have a background in math, I figure that I can more > easily describe my question with a little practical background. I'm > creating a program to find curves; three dimensional curves with > specific endpoints and slope at those endpoints. The flow of the > program will be this: > Input. > 1. Endpoints > 2. Three dimensional unit vector at each endpoint describing the > motion of the curve at that endpoint. > 3. Length of the curve. > Output. > 1. Curve fuction (My biggest problem. I dont know if it is possible to > create one function or colleciton of functions from the limited input. > I've mildly studied polynomial interpolation and splines) If the allowable arc length is greater than the distance between starting point and end point, the curve is possible. If the allowable arc length is equalto the distance between starting point and end point, and the direction at each end is exactly towards the other end, the curve is possible, and will be a line segment. Otherwise the curve is impossible. > Requirements. > 1. Goes through both endpoints > 2. Path of the curve must be along the unit vector at the endpoint > 3. Must have stated curve length. > The optimal curve will : > 1. Be smooth > 2. Minimize concavity I think by concavity you mean what mathematicians would call the curvature > 3. Minimize rate of change of concavity I'm not sure whether you mean to minimize torsion or rate of change of curvature wrt arc length. They are quite different. You might possibly want to minimize both. Both an arc of a circle and a section of a helix have constant positive curvature, but every plane curve has zero torsion, and a helix has constant positive torsion. > (I'm just saying the less curve the better. But the parts that match > up to the unit vector at the endpoints need to be smooth. More broad > curves are better than fewer sharper curves, too.) The Frenet formulas of differential geometry are useful in analysing parametrically represented space curves, including curvature and torsion of such curves. > Can anyone point me in any direction? I'm currently in design time > with the program and I need to find out my options. I dont mind doing > guess-n-check for these values either. So if someone can give me a > template for a curve which fulfills the requirements, I can use trial > and error to find the most optimal of these curves. > I didn't know where to look for help other than here. I hope you all > find enjoyment on helping me with this project. I've talked to people > at my university but those were student tutors, not professors. > Thanks all in advance! > Derek Adams > Student, Computer Science > Cal Poly Pomona Subject: Re: Computer Science student seeks mathematic aid You REALLY need the Mathematics background. The machine can not help you if you lack the Mathematics. Your project absolutely requires you to understand the mathematics. G C >Since I dont have a background in math, I figure that I can more >easily describe my question with a little practical background. >I didn't know where to look for help other than here. I hope you all >find enjoyment on helping me with this project. I've talked to people >at my university but those were student tutors, not professors. > Derek Adams > Student, Computer Science > Cal Poly Pomona Subject: New source of the math!!! I have come across a source providing good support to those experiencing difficulties with doing homework in maths including arithmetic(http://www.bymath.com/studyguide/ari/form1.htm), geometry (http://www.bymath.com/studyguide/geo/pro/pro1/pro1.htm), algebra (http://www.bymath.com/studyguide/alg/pro/pro1/pro1.htm), functions and graphics (http://www.bymath.com/studyguide/fun/pro/pro.htm), principles of analysis(http://www.bymath.com/studyguide/ana/pro/pro1/pro1. htm), all this is also supported by a good deal of examples and illustrations. Subject: Re: Mupad syntax <1Ko3c.1895$Or1.84@news.chello.at > sobald man statt a eine Zahl einsetzt z.B. 2 das Ergeniss nicht richtig > formatiert wird. > also > simplify(2^v/2) Bug: a^v/a is actually a^v*a^(-1), but 2^v/2 is actually 2^v*1/2 (where 1/2 is a rational number of type DOM_RAT) and combine and simplify don't realize this can be simplified. I suggest posting a bug report at www.mupad.de/bugs.html. -- +--+ +--+| |+-|+ Christopher Creutzig (ccr@mupad.de) +--+ Tel.: 05251-60-5525 Subject: Re: Sum evaluation with Wilf-Zeilberger method by support1.mathforum.org (8.11.6/8.11.6/The Math Forum, $Revision: 1.9 primary) id i2CDZe210830; Thank you very much, this was exactly what i needed to prove the Lemma. The function G can be simplified to G = (-1)^(n-k+1)*k*m*binomial(m-k,k)*binomial(m-2*k,n-k+1)/((n+1)* (m-n-1)) Frank Haefner frankdothaefner@web.de Subject: Re: homeomorphisms S^3 --> S^3 Originator: israel@math.ubc.ca (Robert Israel) > it would be great if someone could help me with the following question: > How do I find all isotopy-classes of homeomorphisms S^3 --> S^3? > Is it true, that all homeomorphism, which preserve orientation, are > isotopic to the identity? Yes. Have a look at the archived sci.math.research discussion at http://www.math.niu.edu/~rusin/known-math/99/aut_S3 . Near the bottom of the page, you'll find the following answer by Linus Kramer: > Let Diff(S^3) denote the diffeomorphism group of S^3 (with the > C^infty topolgy), Top(S^3) its homeomorphism group (c-o-topology) > and O(4) the rotation group in R^4. There are natural maps > O(4) --> Diff(S^3) --> Top(S^3). > Hatcher proved [Ann. Math. 117] that both maps are homotopy > equivalences. This settled the so-called Smale conjecture, > and it's a very difficult theorem! Also, the result is not > true any more for spheres of higher dimensions. -- Marc Nardmann (To reply, remove every occurrence of a certain letter from my e-mail address.) Subject: Re: homeomorphisms S^3 --> S^3 Originator: israel@math.ubc.ca (Robert Israel) > it would be great if someone could help me with the following question: > How do I find all isotopy-classes of homeomorphisms S^3 --> S^3? > Is it true, that all homeomorphism, which preserve orientation, are > isotopic to the identity? There have been several replies to this question, mentioning Hatcher's proof of the Smale conjecture. However, an affirmative answer to the original question is older (and is part of background for Hatcher's work). First, that homeomorphisms are isotopic to PL homeomorphisms (and hence to diffeomorphisms) goes back to Moise's work in the late 1950s. (Much of this is reworked in his book, Geometric Topology in Dimension 2 and 3; I think that Shalen's paper: A piecewise-linear method for triangulating $3$-manifolds. Adv. in Math. 52 (1984), no. 1, 34--80, would give a proof more accessible to modern readers.) The proof that an orientation preserving diffeomorphism of the 3-sphere is isotopic to the identity is due to Jean Cerf, from the early 60's: Sur les diffeomorphismes de la sphere de dimension trois $(Gamma sb{4}=0)$. (French) Lecture Notes in Mathematics, No. 53. Daniel Ruberman Subject: Re: Fixed Point Arithmetic Resources ? Originator: israel@math.ubc.ca (Robert Israel) > hi there, > from past few days i have been doing research on using Fixed Point > arithmetic using Integers, i was able to find out some good > resources(google) but they were related to simple > Addition-Subtraction-Multiplication, i m looking for some good resources on > doing Trignometric functions(sin,cos..) through Fixed Point Atithmetic Knuth's METAFONT program implements all the standard trigonometric functions using 32-bit integer arithmetic. It's fairly painful but it works. Subject: Shafer-Dempster vs. Bayes Epigone-thread: zeuwhingri Originator: israel@math.ubc.ca (Robert Israel) Dear Folks, I am currently trying to collect information about the different ways to consider uncertainties in engineering calculations and finally met the Shafer-Dempster Theory of evidence. At some point of the theory there is the conclusion the the Bayes approach is a special case of the Shafer-Dempster theory. I think that I don't understand that. What are the differences of Bayes and Shafer-Dempster (how is Bayes a special case?) and what are the advantages and disadvantages in treating uncertainties? Thanks, Daniel Subject: Re: Pointwise Ergodic Theorem Originator: israel@math.ubc.ca (Robert Israel) Sorry, I was not precise enough: As to the first objection noted I mean that v is the ONLY ergodic measure and as far as the moderators note I intend the map to be nonsingular so that the interval [0,1) cannot map to the point 1. Thanks, E >Suppose you have a map T: X -> X for which the probability measure v >is invariant and ergodic. Also suppose v << m where m is lebesgue >measure. Is it possible to show convergence in the pointwise ergodic >theorem for m almost every x in X. Obviously the ergodic theorem >itself only talks about v almost every convergence. > I don't know anything about this ergodic stuff, but surely the > answer is clearly not? > Maybe X = [0,2], T maps [0,1) to [0,1) and maps [1,2] to [1,2], > and v is concentrated on [1,2]. Then the hypotheses say > nothing whatever about how T behaves on [0,1), so they > certainly don't imply convergence a.e. on [0,1). > [ moderator's note: > Perhaps the OP meant to assume that v is the only ergodic probability > measure. But still there would be easy counterexamples, e.g. > with X = [0,2], where T maps [0,1) to 1, v must be concentrated on [1,2], > and the behaviour of the iterates of points in [0,1) depends on the > iterates of the single point 1, which may be atypical. > - ri > ] >Thanks, >E > ************************ > Subject: Re: Law of large number in a finite set with distance measure Originator: israel@math.ubc.ca (Robert Israel) >>Consider a discrete set X in which a distance d(x,y) that satisfies the >>triangle inequality is defined. >>Suppose there is a probability measure P of X such that the >>mean of P is m. The mean can be defined by, say, >>m = argmin_{x in X} sum_{y in X} d(y, x) P(y) > This is more like the median than the mean. It may not > be unique. Whether d satisfies the triangle inequality > or is even symmetric is not of great importance, although > it might be in looking at the asymptotics if X is > infinite; the triangle inequality would enable a > somewhat easier proof, but not solve all the problems. >>Now, based on P, we get a sample of size D from X, >>D = {x[1], x[2], ..., x[n]}. >>We then find the mean of D. This can be defined by >>hat{m} = argmin_{ x in X } sum_i d( x[i], x ) >>When n tends to infinity, I expect hat{m} tends to m. > Certainly if X is finite, hat{m} will tend to the > set of medians almost surely. Yeh... I also notice about that shortly after I posted the question. >>Is this true? If yes, Can you give me some references that >>contain the proof? If no, is it possible to impose >>additional assumptions on P to make hat{m} go to m? >>Also, can anything be said about >>the rate of convergence, if it can be defined after all? > The probability that hat{m} is not in the set of medians > tends to 0 exponentially in the finite case. If there are > two medians, they are roughly equally likely for large n, > but if there are more, I do not know the asymptotics. This is an interesting property. May I ask where can I find the proofs of these facts? Thank you. Subject: Re: Are analytic Divisors locally algebraic? Originator: israel@math.ubc.ca (Robert Israel) > Let V be an analytic divisor in a polydysc D in C^n (i.e. V={f=0} with > f analytic) > that pass through 0 (i.e. f(0)=0) > Is there (up to shrinking D) a biholomorphism G:D->D, that fixes 0 and > such that the image of V is algebraic (i.e. G(V)={g=0} with g > polynomial) ? > Of course the question is trivial if V is irreducible and smooth at 0. > Thanks in advance for any hint. Your notations are somewhat unclear for me. Do you mean it as analytic germs (with natural structure through f) or as analytic sets (with the reduced structure)? A common formulation would be 'hypersurface singularity' and your Q means to look at it as analytic algebras, asking whether they are 'analytifications' of polynomial ones. For the natural structure i think there are counter examples. Sorry, i do not have one, but suggest to look at ( exp(x) - 1 )^2 - y*exp(x*y)^3 or x^2-y*exp(y*x)^3 (a modification of Neil's parabola x^2-y^3). Playing with Maple it should be singular in 0 (evaluate the Hessian) but to decompose exp(x*y) (in polynomials) should fail. The other suggestion is to search for 'affine GAGA' or similar. May be papers of Greuel or Loijenga are also a good starting point. Or literature on local analytic algebra or geometry (elder ones will be in French). --- remove the no for mail Subject: Re: Are analytic Divisors locally algebraic? Originator: israel@math.ubc.ca (Robert Israel) >Let V be an analytic divisor in a polydysc D in C^n (i.e. V={f=0} with >f analytic) >that pass through 0 (i.e. f(0)=0) >Is there (up to shrinking D) a biholomorphism G:D->D, that fixes 0 and >such that the image of V is algebraic (i.e. G(V)={g=0} with g >polynomial) ? (Probably you mean e.g., not i.e.) >Of course the question is trivial if V is irreducible and smooth at 0. If you really want the biholomorphism to be onto D, then it may not be as trivial as you assert. You should look at the Weierstrass Preparation Theorem; for references see Gunning or Griffiths-Harris. (This assumes that you're willing to accept a biholomorphism from D to an open neighborhood of 0.) Subject: Re: Are analytic Divisors locally algebraic? Originator: israel@math.ubc.ca (Robert Israel) Thanks for your answer. >Let V be an analytic divisor in a polydysc D in C^n (i.e. V={f=0} with >f analytic) >that pass through 0 (i.e. f(0)=0) >Is there (up to shrinking D) a biholomorphism G:D->D, that fixes 0 and >such that the image of V is algebraic (i.e. G(V)={g=0} with g >polynomial) ? > (Probably you mean e.g., not i.e.) I don't see any other definition of algebraic divisor. But anyway it would be good for me if the image of V can be written as zero of a polynomial. >Of course the question is trivial if V is irreducible and smooth at 0. > If you really want the biholomorphism to be onto D, then it may not be > as trivial as you assert. Sorry about that. As you said, it does not need to be onto. On the other hand I don't see how Weierstrass thorem helps. I want g (as above) to be algebraic with respect to all the variables, and induction doesn't seem to work so naturally.