mm-423 === Subject: : Re: The slide rule from hell (found)Johnny:Hey, don't knock em... I grew up using those things... Designed lots of electronic products that sold in the hundreds of thousands of units using nothing but my trusty Hemmi Bamboo log-log rule.These days there seems to be quite a bit of interest from slide rulecollectors who pay relatively big money for slide rules in good condition, especially the more exotic ones.There are Web sites that specialize in rounding up, refurbishing,and reselling good used slide rules.Have a look at: http://www.sliderule.ca/index.shtmlGoogle around and you will find more sites like this...I had a slide rule about as old as the one you saw on eBay, Ikept it in a drawer for decades... then about 10 years ago, whenwe were moving homes, and I was away, my dear wife sold it at her garage sale for $0.50. It cost me $35 originally back when Iwas a student and I now see the same model refurbishedand selling for upwards of $350 - $400! on that famous cylindrical slide rule... those babys commanda good price today!Slide rules will only become more mm-423 -Professional Consultant - Signal Processing and Analog ElectronicsInditic By-the-Sea, FL> I have often wondered what they were like.> The slide rule from hell can be see at:> http://cgi.ebay.com/ws/eBayISAPI.dll?ViewItem&item=3281361903> Why do you suppose people buy slide rules?> Can't add or subtract, three digits accuracy and you track> decimals in your mind.> There is no Underscore in my real addy.=== === Subject: Re: The slide rule from hell (found) > I had a slide rule about as old as the one you saw on eBay, I> kept it in a drawer for decades... then about 10 years ago, when> we were moving homes, and I was away, my dear wife sold> it at her garage sale for $0.50. It cost me $35 originally back> when Iwas a student and I now see the same model refurbished> and selling for upwards of $350 - $400! That's what wives are for. Cf. Aladdin's wife. > on that famous cylindrical slide rule... those babys command> a good price today! This was not a slide rule but a mechanical calculator (hand- powered), if it is the gadget I recall. Sold by Haverhill's. The problem was that I already had a mechanical calculator, as well as an abacus (which I learned to use pretty fast). Oh yes, BTW, I still have my K&E Log-Log Duplex Decitrig 12 slide rule. I should get it out and polish it up one of these days--like me it is a bit stiff in the joints. When my Dad got it for me in 1953, it was by far my proudest possession. I stopped using it on a daily basis when HP came out with the HP-35 scientific calculator.-- ^^^^^^^^^^^^^^^^^^http://galileo.phys.virginia.edu/~jvn/ God is not willing to do everything and thereby take away our free will and that share of glory that rightfully belongs to us. -- N. Machiavelli, The Prince.=== === Subject: : Re: The slide rule from hell (found)That slide rule in SA was fantastic, great true story, something like 600parts, and great accuracy for the time.Wouldn't mind just seeing one, but they seem exceedingly rare now.> Johnny:> Hey, don't knock em... I grew up using those things... Designed> lots of electronic products that sold in the hundreds of thousands of> units using nothing but my trusty Hemmi Bamboo log-log rule.> These days there seems to be quite a bit of interest from slide rulecollectors who pay relatively big money for slide rules in good> condition, especially the more exotic ones.> There are Web sites that specialize in rounding up, refurbishing,> and reselling good used slide rules.> Have a look at: http://www.sliderule.ca/index.shtml> Google around and you will find more sites like this...> I had a slide rule about as old as the one you saw on eBay, I> kept it in a drawer for decades... then about 10 years ago, when> we were moving homes, and I was away, my dear wife sold> it at her garage sale for $0.50. It cost me $35 originally back> when Iwas a student and I now see the same model refurbished> and selling for upwards of $350 - $400!> on that famous cylindrical slide rule... those babys command> a good price today!> Slide rules will only become more valuable to collectors as time> goes on...> --Professional Consultant - Signal Processing and Analog Electronics> Inditic By-the-Sea, FLI have often wondered what they were like.The slide rule from hell can be see at:http://cgi.ebay.com/ws/eBayISAPI.dll?ViewItem&item= 3281361903Why do you suppose people buy slide rules?Can't add or subtract, three digits accuracy and you trackdecimals in your mind. There is no Underscore in my real addy.=== === Subject: : Re: The slide rule from hell (found)Twit,the Empire State Building was designed with one of those.So were all buildings, bridges, electronics, TV and everything else before1970!Including your mom's and grandma's car.And the airplane she flew in on!And the spudnik,What do you think the Germans, English and Americans used during W.W.II?> I have often wondered what they were like.> The slide rule from hell can be see at:> http://cgi.ebay.com/ws/eBayISAPI.dll?ViewItem&item=3281361903> Why do you suppose people buy slide rules?> Can't add or subtract, three digits accuracy and you track> decimals in your mind.> There is no Underscore in my real addy.=== === Subject: : Re: The slide rule from hell (found) I mean to say, Why do people buy them now. I know how touse one and know the history of them. (Invented by Roget of thesaurus fame.)I have a sterling slide rule sitting in a drawer somewhere. The other post explained the collector value of slide rules.I was unaware of this.Twit-ally dee me but not that dumb,-Johnny_[no underscore in my real addy.}> Twit,> the Empire State Building was designed with one of those.> So were all buildings, bridges, electronics, TV and everything else before> 1970!> Including your mom's and grandma's car.> And the airplane she flew in on!> And the spudnik,> What do you think the Germans, English and Americans used during W.W.II?I have often wondered what they were like.The slide rule from hell can be see at:http://cgi.ebay.com/ws/eBayISAPI.dll?ViewItem&item= 3281361903Why do you suppose people buy slide rules?Can't add or subtract, three digits accuracy and you trackdecimals in your mind. There is no Underscore in my real addy.=== === Subject: : Re: The slide rule from hell (found) I mean to say, Why do people buy them now. I know how to> use one and know the history of them. (Invented by Roget of thesaurus fame.)> I have a sterling slide rule sitting in a drawer somewhere.> The other post explained the collector value of slide rules.> I was unaware of this. Actually invented by Napier, IIRC. Napier's Bones they were calledthen.-- ^^^^^^^^^^^^^^^^^^http://galileo.phys.virginia.edu/~jvn/ God is not willing to do everything and thereby take away our free will and that share of glory that rightfully belongs to us. -- N. Machiavelli, The Prince.=== === Subject: : Re: The slide rule from hell (found)>Actually invented by Napier, IIRC. Napier's Bones they were called>then.Napier invented both logarithms and the unrelated Napier's Bones. The slide rule was actually invented slightly later by William Oughtred. Some useful web links: The Oughtred Society is an organization of collectors of slide rules. http://www.oughtred.org/ Here's a page on Napier's Bones and related topic of the Genaille-Lucas Rulers. http://www.nmt.edu/~borchers/napier/napier.html=== === Subject: : Re: The slide rule from hell (found)Dear BrianPicture 13 bottom of the concise enc. of mathematics, Gellert et al, van norstrand, 1977 isbn 0 442 22462 2 from the GDR (old GDR that is)the photo is in tha back of the book and dates (about 1600) it says for Chinese Slide RulePaul>Actually invented by Napier, IIRC. Napier's Bones they were called>then.> Napier invented both logarithms and the unrelated Napier's Bones. > The slide rule was actually invented slightly later by William Oughtred.> Some useful web links:> The Oughtred Society is an organization of collectors of slide rules.> http://www.oughtred.org/> Here's a page on Napier's Bones and related topic of the Genaille-Lucas> Rulers.> http://www.nmt.edu/~borchers/napier/napier.html=== === Subject: : Re: The slide rule from hell (found)Dear BrianI thought it was invented by the Chinese!Paul>Actually invented by Napier, IIRC. Napier's Bones they were called>then.> Napier invented both logarithms and the unrelated Napier's Bones. > The slide rule was actually invented slightly later by William Oughtred.> Some useful web links:> The Oughtred Society is an organization of collectors of slide rules.> http://www.oughtred.org/> Here's a page on Napier's Bones and related topic of the Genaille-Lucas> Rulers.> http://www.nmt.edu/~borchers/napier/napier.html=== === Subject: : svd for complex matrices by support1.mathforum.org (8.11.6/8.11.6/The Math Forum, $Revision: 1.9 primary) id i2F3I0k20075;HI,I'm trying to write a c program that allows me to perform svd of acomplex matrix so that [Uc,Wc,Vc]=svd(Ac), Ac=UWV*. I have a programthat performs svd for real matrix and i know that I must replce everyz=x+iy by a matrix (x -y; y x) so that i can transform my complexmatrix into a real one and let my program give me Ur, Vr and Wr 2m*2nand 2n*2n matrices. But what is the correlation between(Uc,Ur),(Vc,Vr) and (Wc,Wr).Please I need a quick reply.=== === Subject: : Re: svd for complex matrices >HI, >I'm trying to write a c program that allows me to perform svd of a >complex matrix so that [Uc,Wc,Vc]=svd(Ac), Ac=UWV*. I have a program >that performs svd for real matrix and i know that I must replce every >z=x+iy by a matrix (x -y; y x) so that i can transform my complex >matrix into a real one and let my program give me Ur, Vr and Wr 2m*2n >and 2n*2n matrices. But what is the correlation between >(Uc,Ur),(Vc,Vr) and (Wc,Wr). >Please I need a quick reply. >multiply the complex version, writtem in real and imaginary parts out.then you will be able to compare real and imaginary components with the realversion. but it is not a good idea to do this tranformation into a real problemof doubled dimension, since the complete algorithm can be done directly in the complex filed, replacing transposition by taking the complex conjugate transpose.hth=== === Subject: : Re: Kalman filtering by support1.mathforum.org (8.11.6/8.11.6/The Math Forum, $Revision: 1.9 primary) id i2F3IGL20306;I think i can figure out the part of prediction, but on themeasurement part, i am lost again. Isn't prediction and measurementthe same? Since they practically use the same equations (with encodervalues as data) to get the positions and orientation?I am losing sleep and concentration being stuck at this point for sucha long time.jax>3) Can someone guide me to resources where as a beginner i can>learn>more about this modelling of math equation for my scenario?>href=http://www.cs.unc.edu/~welch/kalman/>http:/ /www.cs.unc.edu/~welch/kalman/implementing it.>My X, and Y coordinates are derived thru a few sets of geometric>calculations, namely based on the differences of the encoder values> from the last and current readings. I also have a set ofcalculations>that compute my current orientation to the x-axis. >my process state will be [x, y, orient]transpose. How do i fit this>into a kalman filter? I am truely confused by this, pls advice.=== === Subject: : name for 'element-by-element' productSorry to post out of the blue; I've asked this of a couple of math people,one of whom suggested I try here. Hopefully it's only marginally off topic.I'm documenting a step in an algorithm where we do a pairwise or element-by-element product of two vectors. That is, if we have two rank 1 vectors: A = {a_i} for i elm [0,N) B = {b_i} for i elm [0,N)we form another rank 1 vector of the same size: C = AB = {c_i} by c_i = a_i b_i for each i elm [0,N] (no implied summation)Does this product have a real name? The original developer referred to itas a 'dot product' which is clearly incorrect. Always calling it an'element-by-element multiplication' is a little cumbersome. :)Any advice greatly appreciated, -r=== === Subject: : Re: name for 'element-by-element' product> I'm documenting a step in an algorithm where we do a pairwise or element-> by-element product of two vectors. > Does this product have a real name? It is called the Schur or Hadamard product. It is an important operationin algebraic graph theory.Often, the Hadamard product of two vectors a and b is written as a o b.=== === Subject: : Re: about regularizationDear all, I would like to solve a linear system Ax=b, A is ill-condition. Theconditional number is about 2*10^5. A is upper triangular matrix. I haveadded noise on b so the solution is not good if solving it directly. Iwantto add a regularization operator to improve the solution. For aninstance, Ican add the Laplacian matrix but it is not upper triangular. What theuppertriangular do you suggest me to add as good as the Laplacian matrix.> That seems a mild condition number. In double precision you should> get at least 10 digits of accuracy, in quad 25 digits. Is your data> (A,b) of higher accuracy than that?=== === Subject: : Re: about regularizationIn an ill-conditioned problem with noisy right hand side you cannot expecttoo much accuracy. Read my regularization survey, avaialble from my web site.=== === Subject: : help please by support1.mathforum.org (8.11.6/8.11.6/The Math Forum, $Revision: 1.9 primary) id i2FE56Z19925;iam looking for an efficient method to find thehighest eigen value of a positive semidefinitesymmetric matrix (it may not be sparse). cananybody let me know how to achieve this.=== === Subject: : Re: help please> iam looking for an efficient method to find the> highest eigen value of a positive semidefinite> symmetric matrix (it may not be sparse). can> anybody let me know how to achieve this.If you understand the physics of the system you canguess a trial eigenvector containing a couple of parametersand minimize w.r.t. those. That's usually faster than tryingto minimize the quadratic form w.r.t. n-1 parameters (it isnormalized after all).Faster yet is to guess x_0 and then iterate via ax_n = A x_{n-1}where a is the eigenvalue you seek and A the matrix. You haveto keep normalizing the vectorsat each iteration. Ultimatelyyou will have a approx ( x_n, Ax_n ) / ( x_n, x_n )Note that if the precision of the eigenvector is epsilon, thatof the eigenvalue will be epsilon^2 . So the convergence is usuallypretty rapid.Anyway, the latter is the method I would use (and have used manytimes).-- ^^^^^^^^^^^^^^^^^^http://galileo.phys.virginia.edu/~jvn/ God is not willing to do everything and thereby take away our free will and that share of glory that rightfully belongs to us. -- N. Machiavelli, The Prince.=== === Subject: : Re: help please >iam looking for an efficient method to find the >highest eigen value of a positive semidefinite >symmetric matrix (it may not be sparse). can >anybody let me know how to achieve this. >depends. normally, one would propose Lanczos' method if the matrix is large.(there are ready to use codes in netlib : http://www.netlib.org)but if the dimension is very large, often the simultaneous vector iteration (ritzit) does better. rizit is contained in the svdpack of netlib.if dimension is not large, then a good shift mu (using an _upper_ boundfor the largest eigenvalue and inverse iteration should be faster, but thisrequires the solution of linear systems with the matrix A-mu*I)finally, the maximization of the rayleighquotient x'*A*x on the unit sphere (using conjugate gradient maximization (minimize thenegative) and backprojection on the unit sphere is a simple and oftenefficient method.hth === === Subject: : Re: solve linear systems in boost::numerics::ublas by support1.mathforum.org (8.11.6/8.11.6/The Math Forum, $Revision: 1.9 primary) id i2FEZd524649;ATLAS can be build with msvc, actually this is supported and describedby the current release. I compiled it but have not used the library sofar in msvc projects.blitz++ needs a very advanced (standard conformant) compiler. Maybe itis possible to use it with MSVC 7.1.tnt (jama) should easily compile with most compilers. Which version ofmsvc do you use? Some time ago I compiled it for borland and msvc.Christian=== === Subject: : Re: Modelling dynamics of linked members> In message 12 Mar>I'm looking at a problem of modelling a mechanical system that can be >approximated by a number (say 20) of rigid members (rods), linked to >each other endwise to form a chain. > Have you looked at OpenDE, the Open Dynamics Engine?> This is a library which supports various elements which can linked together.> It provides the solver and all the other gubbins and links with OpenGL.> A chain is modelled in one of the test programs provided, so that might give> you a good idea of how it can be tackled.> Have a look at http://opende.sourceforge.net/Interestingly OpenDE was created by Smith, who was then in the School of Engineering, University of Auckland. I am currently teaching in the School of Engineering.Gib=== === Subject: : solution of non-linear dynamical system? by support1.mathforum.org (8.11.6/8.11.6/The Math Forum, $Revision: 1.9 primary) id i2FFefk00995;greetings!i'm a biologist interested in the dynamical system u[t]=1-exp(u[t-1])starting from u[0]=1, in case it matters. can you help me find a solution for arbitrary t?can you help me find an approximate solution for arbitrary t?=== === Subject: : Re: solution of non-linear dynamical system?>greetings!>i'm a biologist interested in the dynamical system u[t]=1-exp(u[t-1])>starting from u[0]=1, in case it matters. Yes it matters! What would've happened if you'd started with u[0]=0?>can you help me find a solution for arbitrary t?No; there's no easy formula.>can you help me find an approximate solution for arbitrary t?Sure: for large t you'll find u[t] alternates from positive tonegative, that is, you're really interested in the iterationsu[t] = 1 - exp( 1 - exp( u[t-1] ) ). But once u[t] starts to besmall, you can estimate this function with its Taylor series,x - x^3/6 - x^4/24 + ... . (This is enough to show that the u'swill steadily decrease in magnitude once small.) This expansionis close to consistent with the sequence v[t] = sqrt(3/n) andit seems the u's approach these quickly in magnitude.=== === Subject: : Re: solution of non-linear dynamical system? by support1.mathforum.org (8.11.6/8.11.6/The Math Forum, $Revision: 1.9 primary) id i2FNKLp23700;,thanx for your reply :-)i feel really stupid, though :-( after reading your reply, i realizedthat i left a minus sign out. the real equation isu[t]=1-exp(-u[t-1]) , so u is always positive, and decreasing. wouldthat make things any easier?following your lead, i tried to taylor 1 - exp(- (1 - exp(-u)))getting u - u^2 + 5u^3/6 - 5u^4/8 + ... but i have no idea whatsimple function in t would be close to this. i feel really bad about wasting your time on the wrong equation.still, if you or anyone else is still willing to help... i'dappreciate!>greetings!>i'm a biologist interested in the dynamical systemu[t]=1-exp(u[t-1])>starting from u[0]=1, in case it matters. >Yes it matters! What would've happened if you'd started with u[0]=0?>can you help me find a solution for arbitrary t?>No; there's no easy formula.>can you help me find an approximate solution for arbitrary t?>Sure: for large t you'll find u[t] alternates from positive to>negative, that is, you're really interested in the iterations>u[t] = 1 - exp( 1 - exp( u[t-1] ) ). But once u[t] starts to be>small, you can estimate this function with its Taylor series,>x - x^3/6 - x^4/24 + ... . (This is enough to show that the u's>will steadily decrease in magnitude once small.) This expansion>is close to consistent with the sequence v[t] = sqrt(3/n) and>it seems the u's approach these quickly in magnitude.>=== === Subject: : How do i find the equation of this curve by support1.mathforum.org (8.11.6/8.11.6/The Math Forum, $Revision: 1.9 primary) id i2FJwKp31866;Litres 1/4 1/2 3/4 1 1-1/4 1-1/2 Time (minutes) 9.85 7.83 7.52 7.42 7.02 6.98=== === Subject: : Job opening: member of technical staffThe following job is available immediately:Job title: member of technical staff.Job Description:Developing efficient algorithm for electromagnetic field scatteringoffsemiconductor devices. Assist in the enhancement and improvement ofcurrent thin film algorithms.Qualifications:Ph. D in computational physics or numerical analysis or equivalent,with morethan 5 years experience of demonstrated proficiency in solving complexphysicsproblems and numerical programming in C++, Fortran 90/95.Thermawave is located in Fremont, California.Please send resume to:hchu@thermawave.com=== === Subject: : Network Reliability by support1.mathforum.org (8.11.6/8.11.6/The Math Forum, $Revision: 1.9 primary) id i2G0d8332268;Does anyone know of a source for source code incorporating currentstate of the are techniques in computing network system reliability. Irealize that these are NP-hard problems but am looking for code thatimplements the current algorithmic state of the art.Al Myers