I just started working on my master thesis in the theoretical particle physics and my supervisor asked me to start working with FORM, a script interpreter with basic knowledge of lorentz and dirac structures implemented, for the evaluation of some feynman diagrams. As my first glimpse on how FORM looks and works it doesn't seem to be very user friendly and intuitive. So I started to look around a bit and found that sympy, which I know the basics of, also has a module for lorentz indices and dirac matrices.

So before I work more into one of these I thought I would seek some solid advice from someone who maybe has worked with both or was once confronted with the same decision. Has there ever been a benchmark of these and if FORM is the faster option is it worth the inconveniences? Is there maybe a third option considered superior over the two mentioned?

I thought that would be on-topic here, but please correct me if I'm wrong.

  • $\begingroup$ Welcome to SciComp.SE. I have not used FORM, but checking their tutorial, they claim that it is optimized for that class of problems. While general CAS as Maple or Mathematica (or SymPy in this case) are much slower. $\endgroup$
    – nicoguaro
    Feb 23 '16 at 19:26
  • $\begingroup$ @nicoguaro I think that their depiction of other tools may be biased. Yes, Form may be a specialized tool, but that doesn't prevent other algebra systems from inventing algorithms just as or faster than Form for that kind of tasks. And I wouldn't consider SymPy as generalized as Maple and Mathematica, as it also doesn't "try out everything". That's why I thought of a benchmark. Form may also be faster than Sympy because of it being written mostly in C#. But I can't tell $\endgroup$
    – Philipp
    Feb 25 '16 at 14:13
  • $\begingroup$ SymPy is a general computer algebra system, as Maple or Mathematica, and its goal is to be as capable. In general it will be slower than Maple, Mathematica or Maxima. You can try the Sage Benchamrk in both systems for the speed question. $\endgroup$
    – nicoguaro
    Feb 25 '16 at 15:21
  • $\begingroup$ As I'm no expert in neither of the two options, and my algorithms may vary in their effectiveness, I don't think that the benchmark would be very objective and informative. But thanks for your reply! $\endgroup$
    – Philipp
    Feb 25 '16 at 16:10

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.