This is a very fun article which is an interview with a theoretical physicist and computer scientist called Jos Vermasseren about a computer program called FORM. It’s the computer algebra program that powers all the maths they have to do at CERN.
Basically it comes from a lineage of programs made to calculate the normalisation coefficients for various Quantum Field Theory models, the standand model and so on. To make a useful prediction about cross sections or interaction probabilities or so on you have to normalise it, and to do that you have to integrate it squared, at least, and thats a lot of algebra! Too much for a person. So they made a computer do it.
Some of these programs sound amazing. Schoonship was an earlier one that was written by a guy called Martinus Veltman in assembly. I’d love to see that source code…
Eventually a Schoonship user decided that if he wanted all these extra features and other bits, he ought to make his own. Then we got FORM.
Theres some good discussion about science tool-making in that if you want to do make a long-lived tool, and it takes a bunch of time, you need to get results out of it otherwise your career won’t survive. It takes a long time, so you shouldn’t only be writing software. I guess I’ll need to remember that! If you make a good tool, all the citations come a long time later, once you’ve survived your early-career.
FORM is at the heart of advances because its the only thing good and fast enough at the cutting edge of algebra. Mmmm thats so hot how does it workkkk. Gosh. Sorry about that.
Theres lots more algebra its needed for as well. Splitting functions?!
Is being developed by Joshua Davies at University of Liverpool.
Some fun features coming up include useful interface bits, and also floating point coefficients which permits better analyses when it comes to monte carlo integration stability. Noice.
The goal is a place where PhDs and post-docs can contribute high-quality code to FORM knowing that they will be recognised for it, and they won’t be penalised by the loss of time on “results”. The applied physicists have sorted this, where software to do/control experiments is rightly viewed as essential, but the same attitude is lacking in the theoretical world, they say.
I feel like “Travellers Desktop Forecaster” would be fun to write in C…