Re: [Maxima-discuss] Schoonship Form

[Michel T. <ta...@lp...>]

Le 04/12/2022 à 05:34, Andrey G. Grozin a écrit :
> Some statements are still wrong.
>
> On Sat, 3 Dec 2022, Michel Talon wrote:
>> Some statements are wrong, notably Schoonship was developed by 
>> Veltman in assembly language of the
>> CDC7600, a beautiful computer who disappeared long ago. The machine 
>> words were quite long so Veltman used
>> parts of words to store information. With this program he did 
>> computations of radiative corrections in the standard model
>> which gave him access to the Nobel prize. When the CDC disappeared
> Veltman has ported it from CDC assembly language to Atari assembly 
> language
OK, i didn't know, and would have never supposed.
>
>> the program was rewritten by Vermaseren in fortran,
>> and given the name of Form.
> Not in fortran but in C

In fact first in Fortran and then in C.

>> Presently i suppose such extremely complicated computations are no 
>> more of
>> pressing interest to physicists, explaining the disillusion apparent 
>> in the article.
> form is still widely used in theoretical particle physics community

I didn't doubt that, i have done my thesis in this domain. I just said 
that such sort

of "particle physics" is no more mainstream in the theoretical physics 
community.

It has been superseded by all sorts of considerations about string theory,

black holes and information, holography, etc.. which could not care less 
about the fifth decimal

in some process at CERN. And the funding and tenures go there.

>> However saying that Mathematica is now
>> the replacement is perhaps an overstatement.
> Certainly not a replacement. form can (relatively) efficiently work 
> with expressions when the size of a single expression is orders of 
> magnitude larger than the main memory. No other CAS can do it.
>
>> For sure Mathematica is used to draw the Feynman diagrams, but computing
>> them is another story.
> There are important tools for computing Feynman diagrams written in 
> Mathematica, such as
>
> FeynCalc and FormCalc (the last one generated form sources, runs form 
> and reads results back to Mathematica)

I was alluding to that. So the real computation is indeed delegated to 
Form. The other ones i didn't know,

having left the subject decades ago. I have seen russian people in my 
lab doing several loop computations

by hand (!) and other people writing programs to analyze the 
experimental data at CERN, characterize jets,

etc.  Some were working on loop resummation, a la Broadhurst. Alain 
Connes was a fan of that stuff.  But

as far as i remember the practitioners of these radiative correction 
computations were in Germany.


-- 
Michel Talon