mm-1089 === Subject: Algebraic error-correcting coding with MuPAD? Has anybody ever used MuPAD for encoding/decoding of BCH or Reed-Solomon codes? If so, Id be interested in hearing about how you did it. mm-1089 -- For personal replies, remove not from email address. === Subject: Re: Theoretical Background/Literature of a short CAS > > (already posted in sci.math; I hope its posted correct now) > > I want to implement short CAS for calculating and maipulating > Polinominal expressions. The calculation already works, but now Im > interrested in common technologies for parsing and especially > maipulating these expressions - a short CAS. Is there any good > literature of the theoretical background? > > thx Try Mignotte, Mathematics for Computer Algebra. -- Julian V. Noble Professor Emeritus of Physics ^^^^^^^^^^^^^^^^^^ http://galileo.phys.virginia.edu/~jvn/ Science knows only one commandment: contribute to science. -- Bertolt Brecht, Galileo. === Subject: Convert inches to millimeters by support1.mathforum.org (8.11.6/8.11.6/The Math Forum, $Revision: 1.9 primary) with ESMTP id h82Bujt14620 by legacy.mathforum.org (8.11.6/8.11.6/The Math Forum, $Revision: 1.10 $, legacy) id h82Buir27240 Simply multiply inches by 25.4. (25.4 mm in an inch) Formula: mm = inches X 25.4 === Subject: Re: Convert inches to millimeters Butch Ramsey escreveu na mensagem > Simply multiply inches by 25.4. (25.4 mm in an inch) > Formula: mm = inches X 25.4 Or just ask Google ;) --- Outgoing mail is certied Virus Free. Checked by AVG anti-virus system (http://www.grisoft.com). === Subject: Re: Convert inches to millimeters > Simply multiply inches by 25.4. (25.4 mm in an inch) > Formula: mm = inches X 25.4 Wow. Tough stuff, Butch. -- Prof. Roberto Bagnara Computer Science Group Department of Mathematics, University of Parma, Italy http://www.cs.unipr.it/~bagnara/ mailto:bagnara@cs.unipr.it === Subject: help me! Math Cad by support1.mathforum.org (8.11.6/8.11.6/The Math Forum, $Revision: 1.9 primary) with ESMTP id h844g6t23623 by legacy.mathforum.org (8.11.6/8.11.6/The Math Forum, $Revision: 1.10 $, legacy) id h844g6g26933 im new with math cad and i need some help . Your helping will be good for me on the way im study about comp networks and technician. Please tell me what can i do with it and how to use it to solve the === Subject: Re: help me! Math Cad > im new with math cad and i need some help . Your helping will be > good for me on the way im study about comp networks and technician. > Please tell me what can i do with it and how to use it to solve the You should join the Mathcad Collaboratory. You can get there from the Help menu in Mathcad, via the Resource Center. Or, with a browser, http://collab.mathsoft.com/~mathcad2000 === Subject: Trouble importing Mathematica EPS les into Deneba Canvas Hi all, Would send this to comp.soft-sys.math.mathematica, but it is currently down. Maybe someone here has faced a similar problem I am trying to annotate some 2D plots made in Mathematica 5.0 (Mac OS X). I exported the plot as EPS. However, when I import it into Canvas 9 (Mac OS X), a few of the fonts are messed up. For example, what is supposed to be a bold parenthesis ( appears as a radical sign. I cant decide if this is a Mathematica problem or a Canvas problem. A few observations: * Canvas does have access to the Mma-specic fonts, such as Mathematica2. However, when I type ( using the Mathematica2 font within a Canvas text eld, I get the same radical sign which appears when importing the 2D plot. * I tried exporting the EPS le with and without the embedded Mathematica fonts. Still had the same result. * I can convert the EPS le to PDF using ghostscript and it looks great using Adobe Acrobat 6 and Preview.app - all fonts are scalable, not bitmapped. * If I use the tool eps2eps on the EPS le, some of the fonts, e.g. the (, are rasterized while others are preserved as outline fonts. Id rather not use bitmap versions of the fonts as I intend to scale up the plot considerably. (eps2eps is part of the ghostscript collection of tools and will essentially clean up an EPS le) * The same problem occurs when importing the ghostscript-converted PDF le into Canvas. * You can look at the problematic EPS and PDF les here: http://kc144.dyndns.org/~kevin/temp.eps http://kc144.dyndns.org/~kevin/temp.pdf Kevin -- Kevin Gross Ph.D. Student Air Force Institute of Technology Wright-Patterson AFB, Ohio === Subject: Are there any recent reviews of symbolic math packages? packages such as Mathematica, Maple, MuPad, or whatever is in use today? package for genetic programming in it. Not being able to declare the type for a variable and having only one data structure, a list, has driven me to forsake my beloved one. The hoops you have to go through to keep Mathematica from evaluating your expressions while evolving them, makes it difcult to use a number of the symbolic algebra features. The additional overhead I have to have to keep all the expressions unevaluated until the time that I want them evaluated has slowed the system down by a factor of 2.5. I would like to nd a symbolic package with strong typing of variables and structured data types. Furthermore, if I represented the operator evaluation of an equation in a tree form, such as is done in genetic programming, I would like to be able to use the built in algebraic routines to reduce or factor the equations. === Subject: Re: Are there any recent reviews of symbolic math packages? > package for genetic programming in it. [list of troubles...] I sympathsize. I worked on a 50,000(!) line Mathematica program back in the early 90s for a few years, and came to dislike most of what MMa offered me for similar reasons. It was good at math, but not at supporting software engineering. > I would like to nd a symbolic package with strong typing of variables > and structured data types. Furthermore, if I represented the operator > evaluation of an equation in a tree form, such as is done in genetic > programming, I would like to be able to use the built in algebraic > routines to reduce or factor the equations. You might nd the DMS Software Reengineering Toolkit to be interesting. You can dene arbitrary notations (usually programming languages, but you can easily do simpler languages like algebra... weve even fed it a Mathematica syntax denition!) and then apply symbolic rewrite rules written *using that notation* to your hearts content. The rewrite engine understands associative and commutative laws. You can also drop down into procedural code and do arbitrary tree-hacking/procedural transforms. The base language is a compiled parallel programming language (might be a plus for genetic programming), so you dont pay interpretive overhead and you can use those multiple CPUs you can by in cheap workstations. Much of the scale motivation for DMS came from that early bad MMa experience: scale is important. See http://semanticdesigns.com/Products/DMS/DMSToolkit.html. -- Ira D. Baxter, Ph.D., CTO 512-250-1018 Semantic Designs, Inc. www.semdesigns.com === Subject: Re: Are there any recent reviews of symbolic math packages? > > packages such as Mathematica, Maple, MuPad, or whatever is in use today? > > package for genetic programming in it. Not being able to declare the > type for a variable and having only one data structure, a list, has > driven me to forsake my beloved one. The hoops you have to go through to > keep Mathematica from evaluating your expressions while evolving them, > makes it difcult to use a number of the symbolic algebra features. The > additional overhead I have to have to keep all the expressions > unevaluated until the time that I want them evaluated has slowed the > system down by a factor of 2.5. > > I would like to nd a symbolic package with strong typing of variables > and structured data types. Furthermore, if I represented the operator > evaluation of an equation in a tree form, such as is done in genetic > programming, I would like to be able to use the built in algebraic > routines to reduce or factor the equations. > This is not quite the answer you want, but some free advice. Whenever you program anything not already available, using a symbolic math package--it really doesnt matter which--or MathCad or MatLab, it is usually interpreted rather than compiled. This means VERY SLOW. If you need a genetic algorithm for anything, it should be written in a compiled language. I would recommend Forth for ease of programming, debugging and testing, but C or Fortran would do ne. Many modern Forths compile to native code, so are jas fast as C. Even some of the interpretive Forths--GForth, e.g.--are nearly as fast as optimized C. For what it is worth. -- Julian V. Noble Professor Emeritus of Physics ^^^^^^^^^^^^^^^^^^ http://galileo.phys.virginia.edu/~jvn/ Science knows only one commandment: contribute to science. -- Bertolt Brecht, Galileo. === Subject: Re: Are there any recent reviews of symbolic math packages? > > Whenever you program anything not already available, using a > symbolic math package--it really doesnt matter which--or > MathCad or MatLab, it is usually interpreted rather than > compiled. This means VERY SLOW. If you need a genetic algorithm > for anything, it should be written in a compiled language. > Lisp, the implementation language of Maxima, is available in most versions, with a compiler. The usual representation of equations in Lisp would be as a tree. Good luck. === Subject: Re: Are there any recent reviews of symbolic math packages? > > Whenever you program anything not already available, using a > symbolic math package--it really doesnt matter which--or > MathCad or MatLab, it is usually interpreted rather than > compiled. This means VERY SLOW. If you need a genetic algorithm > for anything, it should be written in a compiled language. Lisp, the implementation language of Maxima, is available in most > versions, with a compiler. The usual representation of equations in Lisp > would be as a tree. > > Good luck. I should have remembered that. A modern fast Lisp would probably work. -- Julian V. Noble Professor Emeritus of Physics ^^^^^^^^^^^^^^^^^^ http://galileo.phys.virginia.edu/~jvn/ Science knows only one commandment: contribute to science. -- Bertolt Brecht, Galileo. === Subject: Re: Are there any recent reviews of symbolic math packages? > packages such as Mathematica, Maple, MuPad, or whatever is in use today? > > package for genetic programming in it. Not being able to declare the > type for a variable and having only one data structure, a list, has > driven me to forsake my beloved one. The hoops you have to go through to > keep Mathematica from evaluating your expressions while evolving them, > makes it difcult to use a number of the symbolic algebra features. The > additional overhead I have to have to keep all the expressions > unevaluated until the time that I want them evaluated has slowed the > system down by a factor of 2.5. > > I would like to nd a symbolic package with strong typing of variables > and structured data types. Furthermore, if I represented the operator > evaluation of an equation in a tree form, such as is done in genetic > programming, I would like to be able to use the built in algebraic > routines to reduce or factor the equations. > > Maplesoft has a list of reviews of Maple since January, 2002: -- G. A. Edgar http://www.math.ohio-state.edu/~edgar/ === Subject: Re: Are there any recent reviews of symbolic math packages? It turns out that at least two of these are not genuine reviews at all. Technical Computing Magazine is the brochure of one Maples sales outlets. Its a bit sleazy quoting your own marketing as if it was an independant review. > Maplesoft has a list of reviews of Maple since January, 2002: > === Subject: Re: Are there any recent reviews of symbolic math packages? I would like to thank all the people who provided suggestions. As was pointed out below, many of the people worked for or had an interest in the suggestion they made. I have no objection to a sales or marketing person sending me their pitch, I did make a request for information. Most the suggestions seemed pretty clear as to the company aßiations. To be honest, I did not really pay much attention to this because none of the suggestions addressed the items I am researching. There were a number of suggestions for some low level packages that were suggested as an alternative to using Mathematica. This did not interest me. I like the large symbolic math programs because of all the features that are offered and the fact that I dont have to write the feature. There were some points made on performance issues of using the large symbolic math programs for genetic programming. In fact I raised a point in my original e-mail. My position is that I am willing to take the hit of using a large symbolic math program because of some research questions I want to ask using the computer algebra features. It would certainly take me far longer in programming time to implement a genetic programming package in C++ or C and add the computer algebra features (that worked and were bug free) than implementing a genetic programming package in a symbolic mathematics package. My question is if I might nd a faster symbolic math package that Mathematica to support my research. Let me explain my issue a little further. For a detailed explanation of genetic programming see any book by John Koza, Genetic Programming, Genetic Programming II, or Genetic Programming III. I want to evolve programs that are represented as trees. The mathematical operators are the interior nodes and the variables or constants are the leaves. When you look at an expression in Mathematica using FullForm[], you see that this is the native way that expressions are represented. This is what initially drew me to use Mathematica. Each tree or Mathematica expression is called an individual in genetic programming. A large number of individuals are used in the paradigm, called a population. Each of these individuals has to be stored, normally in an array. In Mathematica it is a list where each element in the list is a Mathematica expression. The problem is that anytime something is stored in Mathematica, Mathematica wants to evaluate it. It wants to put the expression in a canonical form. When it does, it rearranges your tree, or worst it may replace some of the mathematical operations with canonical operations. If you have Mathematica try the following expression, FullForm[Divide[x,y]], to see what is does to a simple operation. This is a really bad feature to have for genetic programming, only the genetic operators should rearrange the trees. There are some Mathematica functions you can use to wrap your tree to delay evaluation, but they dont hold evaluation of everything. What all the previous authors of genetic programming packages in Mathematica have done is use the rewrite facilities of Mathematica to stop the evaluation. During the phases where the trees are being built or manipulated, each mathematical operation is given an alternative name. As there are no functions or features associated with the alternative name, when the tree is automatically evaluated when assigned to its storage place, no changes occur in the tree. When the tree is ofcial evaluated for its value, the rewrite capability is used to replace the alternative name with its real name. For example MyInt -> Int. In my research, I would have 4000 trees with 100,000 nodes. That is a lot of term rewriting. Another issue is that I would like to use the symbolic algebra features on the tree. An example, take some messy sub-tree and simply the expression. Since the trees are stored using alternative function names, the names are basically dummy labels, they dont have the same algebraic properties as the actual Mathematica function. If I gave them the properties of the original functions, I suspect I would get into the original problem of Mathematica wanting to place the expression in canonical form. There is a workaround, bring the tree back to its original representation, algebraically manipulated it, put it back in the alternative form, and store it. First, I want to state that there are work arounds for all the above issues in Mathematica and I can conduct the research I am investigating. The work arounds are additional overhead. I am just trying to see if the grass may be greened with Maple, MuPad, or some other package. > It turns out that at least two of these are not genuine reviews at > all. Technical Computing Magazine is the brochure of one Maples > sales outlets. > > Its a bit sleazy quoting your own marketing as if it was an > independant review. > > > >>Maplesoft has a list of reviews of Maple since January, 2002: >> === Subject: Symbolic math Perl module I am developing a symbolic math module in Perl. It doesnt have the scope of most commercial packages for symbolic calculations. Its aimed at software engineers that need symbolic mathematics, not at mathematicians that engineer software. (Thats a bold statement and not to be taken quite literally, but it gets the point across.) Its capabilities are limited, but its not only open-source, it is also fairly easy to extend. Even from within a program that uses it. I am taling about the Math::Symbolic module the most recent release of which can be found at http://sourceforge.net/projects/math-symbolic/ and as soon as the CPAN indexer picks it up also via http://search.cpan.org/author/SMUELLER/Math-Symbolic/. The most recent release is v0.120. Id appreciate if you had a look at its API, documentation (my English jargon is severely lacking), capabilities and general usefulness (or lack thereof) and offer some feedback. Steffen M.9ler === Subject: what is this group for? sorry but i dont quite get the meaning of math.symbolic === Subject: Re: what is this group for? > sorry but i dont quite get the meaning of math.symbolic > > This newsgroup deals with computer programs for doing calculations symbolically. === Subject: Help in obtaining Macsyma Hello All, I have used Macsyma for many years. However the company went out of business. Yesterday I had a major problem with my PC and was forced to reformat the hard drive and reinstall everything. I *thought* I had my crucial apps and data backed up on oppies, but some of my oppies are giving CRC read errors. Oops. The original distribution media are long gone. Bottom line: my copy of Macsyma is gone. Does anyone know who holds the rights to it? (I know Vaxima is public domain, I want the PC version of Macsyma). How much is an individual copy of Maple? (I wont touch Mathematica). You can lead a horses ass to knowledge, but you cant make him think. > My instructor said the following math answers to my equations used in a SPSS > output were wrong. It seems like basic math to me and my answers keep coming > out the same. Can any of you show me right or wrong? > > A. 2.919 - .377(2) - .026 (6) > I said 2.321; he said 2.009 Hes right: 0.377 * 2 = 0.754 0.026 * 6 = 0.156 2.919 - 0.754 - 0.156 = 2.009 > B. 2.919 - .377(3) - .026 (14) > I said 1.495; he said 1.424 Hes right: 0.377 * 3 = 1.131 0.260 * 14 = 0.364 2.919 - 1.131 - 0.364 = 1.424 > C. 2.919 - .377(4) - .026 (3) > I said 1.586; he said 1.33 Hes right: 0.377 * 4 = 1.508 0.036 * 3 = 0.078 2.919 - 1.508 - 0.078 = 1.333 -- Rich Carreiro rlcarr@animato.arlington.ma.us === Subject: Re: Help, please~ > Can anybody help me to prove following ? > When G=G_0 > G_1 > ... > G_n={e} be an abellian subnormal series of a > group G, > If N is a normal subgroup of a group G, > Prove : G_(i+1) *N is a normal subgroup of G_i *N (i=0,1,2,...,n-1) > For x in G_i *N , y in G_(i+1) *N, > If I can show x*y*x^-1 is in G_(i+1) *N then this proof will be complete. Well, you must also show that {G_j N} is a subgroup... > But, I dont know how can I show above. > If someone can help me please post reply. Heres how to show that {x y x^-1} is in {G_{i+1} N}: Let H = G_i and K = G_{i+1}. Then x = {h n} and y = {k n} for some h in H, k in K, n and n in N, so {x y x^-1} = (h n) k n (n^-1 h^-1) = (n h) k n (h^-1 n^-1) (since N is a normal subgroup of G) = n h k (n h^-1) n^-1 = n h k (h^-1n) n^-1 (again since N is normal) = n h k h^-1 (n n^-1) = n h k h^-1 n (since N is a subgroup) = n (h k h^-1) n = n k n (since K is normal in H) = (n k) n = (k n) n (again since N is normal) = k (n n) = k n (again since N is a subgroup) is in {K N}. I dont know why the problem specied that the subnormal series be abelian, since that isnt used. -- Jim Heckman === Subject: Re: Help, please~ > > Can anybody help me to prove following ? > > > When G=G_0 > G_1 > ... > G_n={e} be an abellian subnormal series of a group G, > If N is a normal subgroup of a group G, > > Prove : G_(i+1) *N is a normal subgroup of G_i *N (i=0,1,2,...,n-1) > > For x in G_i *N , y in G_(i+1) *N, > If I can show x*y*x^-1 is in G_(i+1) *N then this proof will be complete. > But, I dont know how can I show above. > > If someone can help me please post reply. It seems that youve got a lot more hypotheses that you need. Suppose H, K and N are subgroups of G with N normal in G and H ia normal subgroup of K, then H*N is normal in K*N. Notation: h1 = h_1 etc.; h = h^(-1) etc.; hs, ks and ns are elements of the obvious subgroups. To show H*N normal in K*N show (k*n)*H*N*(k*n) is contained in H*N. k*n*h*n1*n*k = k*n*h*k*k*n1*n*k = [N is normal in G] k*n*h*k*n2 = k*n*k*k*h*k*n2 = [H is normal in K] k*n*k*h1*n2 = [N is normal in G] n3*h1*n2 = [N is normal in G so N*h1 = h1*N] h1*n4*n2 which is in H*N Check this carefully - it was done quickly. -- Paul Sperry Columbia, SC (USA) === Subject: df/dx+2f, Been a while since I had calculus Been a while since I had calc. I can do very basic differentiation and integration. Im hard-pressed to nd 2 examples for this... In linear algebra, a problem asks to prove/disprove that the following is a vector space: The set of functions {f: R-->R | df/dx + 2f = 0} and set of functions {f: R-->R | df/dx + 2f = 1} Our teacher told us the rst example is a vector space. The 2nd example is not a vector space due to the fact that the zero vector is not allowed in the 2nd example. He said he would not test us on the above so he did not go and explain what it meant. But Im trying this on my own. Does the notation df/dx + 2f = 0 mean the the derivative of the function(that returns 2*x) plus the output of the function = 0? If so, can someone give me just 2 examples of such a function? I was thinking f(x)=2x and f(x) = x+x but they dont satisfy the df/dx + 2f = 0. How do you nd such functions again? Same with df/dx + 2f = 1. I looked in an old calculus book I have and nothing really jumped out. === Subject: Re: df/dx+2f, Been a while since I had calculus > Been a while since I had calc. I can do very basic differentiation and > integration. Im hard-pressed to nd 2 examples for this... > > In linear algebra, a problem asks to prove/disprove that the following > is a vector space: > > The > set of functions {f: R-->R | df/dx + 2f = 0} > and > set of functions {f: R-->R | df/dx + 2f = 1} > You are making this way too hard. All you need to know is D(f + g) = D(f) + D(g), D(r*f) = r*D(f) and, I guess, D(0) = 0. [D(f) = derivative of f.] Now check the theorem that says A subset of a vector space is a subspace if and only if ... -- Paul Sperry Columbia, SC (USA) === Subject: Re: df/dx+2f, Been a while since I had calculus > Been a while since I had calc. I can do very basic differentiation and > integration. Im hard-pressed to nd 2 examples for this... > In linear algebra, a problem asks to prove/disprove that the following > is a vector space: > The > set of functions {f: R-->R | df/dx + 2f = 0} > and > set of functions {f: R-->R | df/dx + 2f = 1} > Our teacher told us the rst example is a vector space. The 2nd > example is not a vector space due to the fact that the zero vector is > not allowed in the 2nd example. He said he would not test us on the > above so he did not go and explain what it meant. But Im trying this > on my own. > Does the notation df/dx + 2f = 0 mean the the derivative of the > function(that returns 2*x) plus the output of the function = 0? > If so, can someone give me just 2 examples of such a function? I was > thinking f(x)=2x and f(x) = x+x but they dont satisfy the df/dx + 2f = > 0. > How do you nd such functions again? Same with df/dx + 2f = 1. > I looked in an old calculus book I have and nothing really jumped out. df/dx+2*f=0 df/dx=-2*f df/f=-2dx Integrate both sides: ln[f]=-2*x+C f=C*e^(-2*x) Where C, C are arbitrary constants; (C=ln(C) This the general family of solutions. Looking at df/dx+2*f=1 A particular solution is f=1/2 The general solution is f=1/2+C*e^(-2*x) Hope this helps. -- -- Geo. Michael Henry No! Bad dog! I said sit! anonymous === Subject: Re: Basic Math, Help Me Please? <45KdnZp72qk-DfDd4p2dnA@netdoor.com> multiplication and powers? For the same reason my slide rule has separate scales for them... -- P.A.C. Smith This post paid for by the Committee to Re-educate the President. === Subject: Re: Basic Math, Help Me Please? Can you tell me what keys I press on my calculator to get the following: 2.919 ^ -2.321 = 0.0832 I have the TI-30x calculator, which has an e^x symbol right above the LN key. I know I will have to press the 2nd key to get the e^x key to work. I also have a key that says LOG. > https://www.cotse.net/users/dns123/logit.htm The link above goes to 3 problems that I worked. There is a formula > involved. They are very short and should not take much time. Will someone > please check my answers? I am told that all 3 of my answers are wrong, but I > was not told why they were wrong. > First, you are using the wrong value of e. Check out > ; it includes a lot of information about e, > including its numerical value (approximately 2.71828, not 2.919). > Second, you are confusing exponentiation with multiplication. x with superscript > y does not mean x multiplied by y, it means x to the power of y. When writing on > the Internet, x with superscript y is often written as x^y instead. (The caret > symbol points up, indicating thats where the y would be written if we > in the superscript.) > In x^y, x is called the base and y is called the exponent. The value of x^y is > related to multiplication, even though its not as simple as just > multiplication. Specically, when the exponent (y) is a positive integer, x^y > means x multiplied by itself, y times. For example: > 2^3 = 2 multiplied by 2, three times) = 2 * 2 * 2 = 8 > 5^4 = 5 multiplied by 5, four times = 5 * 5 * 5 * 5 = 625 > (-1.5)^3 = -1.5 multiplied by -1.5, three times = (-1.5) * (-1.5) * (-1.5) > = -3/375 > When y is not a positive integer things get more complicated, and I dont have > the time right now to go into all of it here (but I can nd the time later if > you want to hear more about it). > I am surprised you are being asked to compute e^(-z) without being explained > what e is and what x^y means. However, I can say that computing x^y by hand is > impractical in most cases when y is not an integer; its almost always done on a > calculator or a computer. If all you care about is the value (as opposed to > wanting to know why the value is what it is), then all you need to know that to > compute e^(-z), you enter -z on your calculator and push the key labeled e^x > or exp. > hth > meeroh > -- > If this message helped you, consider buying an item > from my wish list: === Subject: Re: Basic Math, Help Me Please? > > Can you tell me what keys I press on my calculator to get the following: > 2.919 ^ -2.321 = 0.0832 > > I have the TI-30x calculator, which has an e^x symbol right above the LN > key. I know I will have to press the 2nd key to get the e^x key to work. I > also have a key that says LOG. One way to do it is to use the relation that a^b = e^(b*log(a)), where log is the natural logarithm. If you have a LN key then LN is the key for the natural log operation. If you dont have a LN key but do have a LOG key, try LOG(e^(1)). If that evaluates to 1, then LOG is your key for the natural logarithm. === Subject: Re: Basic Math, Help Me Please? What I did was: 2.919 [x^y] (thats a key on my calculator) 2.321 [+/-] (thats the change sign key) [=] and 0.0832 popped out. What you probably really wanted to do was 1/(1+e^-2.321), which on most calculators is calculated as: 2.321 [+/-] [e^x] + 1 = [1/x], or possibly as: 1/(1+2.321 [+/-] [e^x])= (that way you get to use parentheses to make it a little more natural). A few calculators (like the Windoze calculator) dont seem to have an e^x key. The trick to get around this is to use inverse functions: 2.321 [+/-] [inv] [ln] + 1 = [1/x]. Oh, yeah, the answer is about .9106. Can you tell me what keys I press on my calculator to get the following: > 2.919 ^ -2.321 = 0.0832 I have the TI-30x calculator, which has an e^x symbol right above the LN > key. I know I will have to press the 2nd key to get the e^x key to work. I > also have a key that says LOG. > One way to do it is to use the relation that a^b = e^(b*log(a)), where > log is the natural logarithm. > If you have a LN key then LN is the key for the natural log operation. > If you dont have a LN key but do have a LOG key, try LOG(e^(1)). > If that evaluates to 1, then LOG is your key for the natural logarithm. === Subject: Re: Basic Math, Help Me Please? My calculator has a e^x key, a LN key, and a LOG key. I have to use the 2nd key to get the e^x key to work. I also have a y^x key, but I dont know what that is for. However, I never can seem to get the calculations to compute equations involving e^x, so I must be doing something wrong. The book does not explain how to enter the following equation into my calculator: Prob = 1 / 1 + e^-z Problem A: 1/1+2.919^-2.321 Can you tell me what keys I press on my calculator to get the following: > 2.919 ^ -2.321 = 0.0832 I have the TI-30x calculator, which has an e^x symbol right above the LN > key. I know I will have to press the 2nd key to get the e^x key to work. I > also have a key that says LOG. > One way to do it is to use the relation that a^b = e^(b*log(a)), where > log is the natural logarithm. > If you have a LN key then LN is the key for the natural log operation. > If you dont have a LN key but do have a LOG key, try LOG(e^(1)). > If that evaluates to 1, then LOG is your key for the natural logarithm. === Subject: Re: Basic Math, Help Me Please? > > Can you tell me what keys I press on my calculator to get the following: > 2.919 ^ -2.321 = 0.0832 > > I have the TI-30x calculator, which has an e^x symbol right above the LN > key. I know I will have to press the 2nd key to get the e^x key to work. I > also have a key that says LOG. I dont know the TI-30x, does it have an x^y key? or just an e^x key? meeroh -- If this message helped you, consider buying an item from my wish list: === Subject: Re: Basic Math, Help Me Please? What I did was: 2.919 [x^y] (thats a key on my calculator) 2.321 [+/-] (thats the change sign key) [=] and 0.0832 popped out. What you probably really wanted to do was 1/(1+e^-2.321), which on most calculators is calculated as: 2.321 [+/-] [e^x] + 1 = [1/x], or possibly as: 1/(1+2.321 [+/-] [e^x])= (that way you get to use parentheses to make it a little more natural). A few calculators (like the Windoze calculator) dont seem to have an e^x key. The trick to get around this is to use inverse functions: 2.321 [+/-] [inv] [ln] + 1 = [1/x]. Oh, yeah, the answer is about .9106. Can you tell me what keys I press on my calculator to get the following: > 2.919 ^ -2.321 = 0.0832 I have the TI-30x calculator, which has an e^x symbol right above the LN > key. I know I will have to press the 2nd key to get the e^x key to work. I > also have a key that says LOG. > I dont know the TI-30x, does it have an x^y key? or just an e^x key? > meeroh > -- > If this message helped you, consider buying an item > from my wish list: === Subject: Re: Basic Math, Help Me Please? > > What I did was: > 2.919 [x^y] (thats a key on my calculator) 2.321 [+/-] (thats the change > sign key) [=] and 0.0832 popped out. Yeah, I wasnt sure if your calculator had that key. > What you probably really wanted to do was > 1/(1+e^-2.321), which on most calculators is calculated as: > 2.321 [+/-] [e^x] + 1 = [1/x], or possibly as: > 1/(1+2.321 [+/-] [e^x])= (that way you get to use parentheses to make it a > little more natural). That works too, and is probably what whoever gave you the problems had in mind. By the way, when replying to a message you shouldnt put all of your reply rst and then all of the original message below it. Its generally preferable to only include the relevant portions of the original message, and write your replies below the appropriate portions of the original message, as I did here. This makes it easier to keep track of what has been said and what you are referring to, especially for those people who arent paying attention to every message that goes by. Hope this helps, meeroh -- If this message helped you, consider buying an item from my wish list: === Subject: Re: Basic Math, Help Me Please? My calculator has a e^x key, a LN key, and a LOG key. I have to use the 2nd key to get the e^x key to work. I also have a y^x key, but I dont know what that is for. However, I never can seem to get the calculations to compute equations involving e^x, so I must be doing something wrong. The book does not explain how to enter the following equation into my calculator: Prob = 1 / 1 + e^-z Problem A: 1/1+2.919^-2.321 Can you tell me what keys I press on my calculator to get the following: > 2.919 ^ -2.321 = 0.0832 I have the TI-30x calculator, which has an e^x symbol right above the LN > key. I know I will have to press the 2nd key to get the e^x key to work. I > also have a key that says LOG. > I dont know the TI-30x, does it have an x^y key? or just an e^x key? > meeroh > -- > If this message helped you, consider buying an item > from my wish list: === Subject: Re: Basic Math, Help Me Please? alt.math.undergrad: >My calculator has a e^x key, a LN key, and a LOG key. I have to use the 2nd >key to get the e^x key to work. I also have a y^x key, but I dont know what >that is for. However, I never can seem to get the calculations to compute >equations involving e^x, so I must be doing something wrong. Not only are you posting upside down, youve posted your last two Cut it out. -- Stan Brown, Oak Road Systems, Cortland County, New York, USA http://OakRoadSystems.com An expense does not have to be required to be considered necessary. -- IRS Form 1040 line 23 instructions === Subject: Re: Basic Math, Help Me Please? > Not only are you posting upside down, youve posted your last two > > Cut it out. You have as much to learn about tact as she has about properly posting on Usenet. I understand what your beef is with proper posting, but nobody will listen to you if you come across as an asshole (and you do, whether you intend it or not). Your tactless postings are not improving the posting quality on the newsgroup by educating people what they should do better, but by discouraging newbies from posting in the rst place. I consider that at least as undesirable as bad attributions, and therefore I would like to urge you to work on how you word your netiquette complaints to newbies. If you dont think you have the time to write out a polite response every time, I suggest you build a boilerplate response which refers them to a suitable webpage. There is no shortage of netiquette guides that are both more useful and more pleasant than the single bitter sentence you usually provide. meeroh -- If this message helped you, consider buying an item from my wish list: === Subject: Re: using GSL > May be it is not quite the right NG, but i try ... > > I have some nal problems in using GSL with MSVC 6 > on Windows: installing 1.3 was ok, demo workspace was > ok as well (with desired output) for the enclosed msvc > example. > > After copying the demo project to my standard and re- > building it agian worked. > > Now i wanted a free demo2 simply using Bessel ftcs > and to create a DLL (with __stdcall and export as the > desired result). > > After some adjustments (according to the readme le > for msvc) [and some failures of mine of course] it does > produce a debug version with a DLL which as desired can > be used (by Excel as user interface) and also gives the > expected results (with almost expected exactness ~ 5e-14 > for x close to 0 and better for 0.7 < x as tested in > Maple [it has the same calling convention]). > > But if i want to have a release version i get an error > unresolved external symbol _gsl_sf_bessel_K1 which may > be the cryptic moaning of MS that it can not nd the > header le (?, i am not very used to C programming). > > Any ideas, suggestions or links to an appropriate source > for resolving my Q? > > Thx > Axel > > --- > remove the no for mail What the heck is GSL? That sounds like a linker error. If so, the debug version is seeing an object or library le with this function or variable, and the release version isnt. Are you getting the error when you try to build, or when you try to run? If its during build, do you have a function or variable in a C le called gsl_sf_bessel_K1? Is the corresponding object le in the dependency list for the release version? -- Tim Wescott Wescott Design Services http://www.wescottdesign.com === Subject: Re: using GSL > What the heck is GSL? http://sources.redhat.com/gsl/ As Googling gsl would have told you, GSL is the GNU Scientic Library, a comprehensive math library written in C. === Subject: Re: FMM to solve Helmholtz equation in the complex case > 4) Is it possible to have a detailed example of the use of the [FMM] > related to the solution of the Helmholtz equation ? The FMM Toolbox http://www.madmaxoptics.com/technology/products/FMMToolbox.asp for Matlab solves PDEs including the Helmholtz using the Fast Multipole Method. It appears to be free with registration. === Subject: Re: Dealing with variable diffusion coefcients and an advection term > > I am trying to solve a diffusion problem with a variable diffusion > coefcient D(x). I have found that in 2D I get signicant > differences when I solve the problem using a O(h^2), ux conservative > method > > df/dt = d/dx(D(x)df/dx) + ... > > or an O(h) upwinding method with the equation in the form. > > df/dt = D(x)*f + dD/dx*f + ... > > I have looked at the papers referenced in Numerical Recipes to nd > higher order methods for solving this upwinding version, but they all > assume the advection term looks like -d/dx(Vf). If I dene V=-dD/dx > then mine is of the form -V(x)*df/dx which would be ne if V were > constant. Can anyone tell me where I can nd a good higher order > advection scheme which assumes the term looks like -V(x)*df/dx? > > Shawn If your actual problem can be expressed as a variable-coefcient diffusion problem: df/dt = -dQ/dx , Q = -D*df/dx Then you should use the conservative nite-difference form, i.e. dQ/dx = ( Q(i+0.5)-Q(i-0.5) )/dx , Q(i) = -D(i)*(f(i+0.5)-f(i-0.5))/dx --> df/dt = -dQ/dx = ( D(i+0.5)*(f(i+1)-f(i)) - D(i-0.5)*(f(i)-f(i-1) )/dx^2 where the diffusivities at mid-points are evaluated using some interpolation, i.e. linear: D(i+0.5) = 0.5*(D(i)+D(i+1)) If your problem cannot be expressed as df/dt = d(D*df/dx)/dx, but is actually an advection diffusion problem: df/dt = D*df/dx^2 + V*df/dx then a good method is given here http://www.ldeo.columbia.edu/~mspieg/ check out the pre-print (available in PDF) A semi-lagrangian crank-nicholson algorithm for the numerical solution of advection-diffusion problems. This is a robust and fairly simple method, applicable to a wide range of Peclet Numbers (i.e. L*V/D, with L a characteristic length, i.e. domain size), which is the ratio of diffusion velocity to advective velocity. It should maintain 2nd order as well. Hope this helps, Matt === Subject: Re: Vector Crossproduct in polar coordinates >I know you cant just do this by stufng the vectors elements into the >Cartesian cross-product formula, and Im also curious if theres a way >to do it that doesnt involve going all the way back to Cartesian >coordinates< Exactly!! That is my point. How does one perform the crossproduct of a 3d polar coordinate pair of vectors in the polar coordinate system WITHOUT using a conversion to Cartesian coordinates. In reviewing a dozen or so calculus or vector analysis texts and Google, I have found no information GC === Subject: Re: Vector Crossproduct in polar coordinates >I know you cant just do this by stufng the vectors elements into the >Cartesian cross-product formula, and Im also curious if theres a way >to do it that doesnt involve going all the way back to Cartesian >coordinates< > Exactly!! That is my point. How does one perform the crossproduct of a 3d > polar coordinate pair of vectors in the polar coordinate system WITHOUT using > a conversion to Cartesian coordinates. In reviewing a dozen or so calculus or > vector analysis texts and Google, I have found no information > GC As GORDO said you do it in exactly the same way as for rectangular coordinates. Use the determinant method with the top row (er, etheta, ez) as unit vectors instead of (i, j, k). The second row is (ar, atheta, az) for vector a and the third row is (br, btheta, bz) for vector b. This gives a X b where a and b are the components of the vectors in polar coordinates and the cross product is also in polar coordinates. Ditto for spherical coordinates. === Subject: Re: Vector Crossproduct in polar coordinates >I know you cant just do this by stufng the vectors elements into the >Cartesian cross-product formula, and Im also curious if theres a way >to do it that doesnt involve going all the way back to Cartesian >coordinates< > > Exactly!! That is my point. How does one perform the crossproduct of a 3d > polar coordinate pair of vectors in the polar coordinate system WITHOUT using > a conversion to Cartesian coordinates. In reviewing a dozen or so calculus or > vector analysis texts and Google, I have found no information > GC Food for thought with spherical coordinates, check into spherical trig or spherical astronomy, this looks solvable via spherical trig and a couple of rotations of the result. The formulae may require more computational work than just biting the bullet and using cartesian coordinates and conversions... a more general problem probably is doable via tensors and the metric tensor, but the details escape me. is not the 3d vector product related to some tensor product. [at least for spherical, and cylindrical coordinates,the metric is diagonal.:)] === Subject: Re: C or C++ routien for calc matrix inverse > the le you posted on vibrationdata cannot be opened without a > subscription...how can i get it without subscribing??? urgent i need > the code for the inverse of a matrix!! help! Get CLAPACK from www.netlib.org. Solve the system A*X=I, where I is the identity matrix. If A is just some general square matrix (no special structure), then use the XGESV routines from CLAPACK, where X is S, C, D or Z (single, single complex, double, double complex). If A has special structure (e.g., triangular, symmetric) then use the Level 3 BLAS (cblas). More important: do you REALLY need an explicit inverse, or are you just trying to solve a system of equations A*X=B? If youre solving a system of equations, dont compute the inverse of A and then multiply by B. Use XGESV and directly solve A*X=B. === Subject: Re: [Introduction fea in Mechanical elasticity problems 2D/3D] So, I will start with this reference. hackervalley === Subject: Collection of IT e-books, IT theses by support1.mathforum.org (8.11.6/8.11.6/The Math Forum, $Revision: 1.9 primary) id i3ABIRm06671; A lot of IT e-books, IT these about DotNet, Oracle, Cisco, Microsoft, Sybex, and more at http://www.diachiweb.com === Subject: Re: Matrix algebra question > > Does anyone know what the condition for the following matrix product to > equal the identity matrix: > > A((AA)^-1)A > > Im nding this often is true - but is it aways true (provided (AA) is > non-singular)? > > > > Mike Let me take a crack at it: Being (AB)^-1 = (B^-1)(A^-1) Then: A((AA)^-1)A = A(A^1)(A^-1)A So, having in mind that : A(A^-1) = I Then it leads to: A(A^1)(A^-1)A = I I = I So, as far as I can tell, the only conditions A and A must follow is that they must allow inversion (non-singular square matrix) and that A and A should be multipliable (compatible dimensions). All in all, A and A should be non-singular, square and with the same dimension. Therefore it seems that yes, it is always true (provided (AA) is non-singular). Please correct me if Im wrong. Hope this helps Rui Maciel P.S.: Forgive my poor english