On 25 Oct 2014 01:55, "Jason Grout" <jason-s...@creativetrax.com> wrote: > > The AMS Notices has a column about using computers to do math, dwelling on some problems they had with Mathematica: > > http://www.ams.org/notices/201410/rnoti-p1249.pdf
Somewhat related I see an example last week of where I think the use of computer software can be a problem. I was at a meeting organised by Keysight (formally Agilent) on THz technology. There was a presentation about the verification of vector network analyzers at THz frequencies by someone working at the National Physics Laboratory (NPL) in the UK - the UK equivalent of NIST in the USA. He had essentially created a verification device out of hollow metallic waveguide. The electrcal properties of the waveguide were computed using CST Microwave Studio, which is a 3D electromagnetic simulation program costing well over 10x that of Mathematica. In order to compute the uncertainty in the electrical properties of the waveguide things like the surface roughness of the metal, uncertainty in the length, width, height etc. I know from experience that there are many similar packages to CST. They all basically compute numerical solutions to Maxwell's Equations, but don't use identical methods. Some for example compute properties using the Finite Difference Time Domain (FDTD) method, then use an FFT to get frequency domain data. Other programs use finite elements and work in the frequency domain. The results of all the packages generally agree well, but no two would give exactly the same result. But when I asked him about the effect of the inexactness of the software, he admitted that there was no uncertainty assumed for the EM simulator. He said he could repeat with other simulation software to get an idea of its effect, and hopefully develop an analytical solution. Whilst I expect it would be possible to develop an analytical solution assuming square walls on the waveguide, I doubt one could do it taking into account all the mechanical errors in making these. So this was a national standards laboratory developing a device to measure the uncertainty in measurements of state of the art instruments, but the software was assumed to be perfect. I would add this device is not a national standard - he was only reporting on work taking place at NPL. Another thing related to the use of 3D electromagnetic simulation software is the huge number of scientific publications making use of such software. Realistically nobody could repeat the work. The authors don't make the files available, and even if they did, some of these programs cost over $100,000. They make the cost off Mathematica look like a childs pocket money. Dr David Kirkby Managing Director Kirkby Microwave Ltd Registered office: Stokes Hall Lodge, Burnham Rd, Chelmsford, Essex, CM3 6DT, United Kingdom Registered in England and Wales as company number 08914892 http://www.kirkbymicrowave.co.uk/ Tel 07910 441670 / +44 7910 441670 (0900-2100 GMT) -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+unsubscr...@googlegroups.com. To post to this group, send email to sage-devel@googlegroups.com. Visit this group at http://groups.google.com/group/sage-devel. For more options, visit https://groups.google.com/d/optout.
