2 ms·
Our Obligation to Build Open Source Software for Modern Science
- MichaelCrawford 12y agoI wrote some high energy physics data analysis software in FORTRAN for my UCSC senior thesis in 1993. I even got to spend the summer at CERN. :-D I required seven weeks to come to grips with CERNLIB and its associated tools such as Patchy, a cross between IBM's JCL and Steve Martin's Cruel Shoes, four days to write my code, then three days to run my simulation long enough to write my thesis. I had dropped out of school to work as a coder for six years, but decided to complete my degree so I'd stop having to explain to why I never graduated in job interviewers. So I knew quite a lot more about software development than even the postdocs. The grad students and postdocs were all stunned at the simple, pristine readibility of my source. I scolded them all with "The reason that physics software is so hard to write, is that YOU PHYSICISTS MAKE IT HARD!" Every law of physics other than general relativity - gravity, loosely speaking - is in CERNLIB somewhere, yet despite spending decades wandering its Gordian labyrinth, I have yet to actually find any.