A colleague of mine talked about a calculation he had pushed to a new height of precision. He had a tool. A computer program from the 1980's.

Particle physicists use a lot of long equations in their work. They draw thousands of pictures called Feynman diagrams that depict possible collision outcomes in order to find signs of new elementary particles. Adding these formulas with computers is a challenge. The particle physics rules we learn in school are not very efficient.

The programs aim to handle these tasks. FORM is the only program that has solved the biggest equations in the world for 33 years.

FORM is a part of the infrastructure of particle physics that is needed for the hardest calculations. As with many essential pieces of digital infrastructure, FORM's maintenance is done by one person. He has begun to take a break from FORM development. The incentive structure of academia has made no successor emerge. Particle physics may be forced to slow down if the situation isn't changed.

When the role of computers was changing rapidly in the 1980's, FORM started. Schoonschip was released as a special chip that you plugged into the side of an Atari computer. The program could be downloaded by universities all over the world. He used FORTRAN to program it. FORM was a reference to that. The programming language he used was called C. His software came out in 1989. The number of institutions that had downloaded it kept increasing.

Every few days, a particle physics paper that cites FORM has been published. The majority of the high-precision results that our group obtained in the past 20 years were based on FORM code.

A procedure for rearranging equations to have as few multiplications and additions as possible was one of the reasons FORM was popular. The most powerful advantage of FORM is its ability to handle memory.

Humans and computers have two different types of memory, main and external. Your computer's main memory is easy to access, but limited in size. Hard disks and solid-state drives are slower than external memory devices. If you want to solve a long equation, you need to store it in main memory.

Both types of memory were limited in the '80s. FORM was built in a time when there wasn't a lot of memory or disk space. The equations were too long for main memory to process. The hard disk needed to be treated as main memory in order to calculate one.

The comic shows the situation well.

Illustration: xkcd.com

FORM uses its own techniques. Each term is assigned a fixed amount of space on the hard disk by the program. This technique makes it easier for the software to keep track of the equation's location. It's easy to bring those pieces back to main memory when you need them.

From 128 kilobytes of RAM in the Atari 130XE in 1985 to 128 gigabyte of RAM in my souped-up desktop, memory has grown a million times. The tricks that were developed remain important. Their need for precision and the length of their equations increases as they search for evidence of new particles in large amounts of data.