Hacker News Reader

Comment by kragen | Hacker News Reader

It's not hard to see from the first couple of chapters why DEC PHIGS and PEX were routed by GL. I mean, the main factor was that Jim Clark took a VLSI design class from Carver Mead and Lynn Conway, doing a 3-D accelerator, and founded SGI with it in 1982: https://ai.eecs.umich.edu/people/conway/Impact/FundingaRevol...

But really, Fortran for examples? In a 1989 book?

View on HN

It wouldn't surprise me if most people using PHIGS/PEX in 1989 were doing visualization for scientific computing on workstations, and any UI or graphics was an afterthought compared to the number crunching code that was generating the data to visualize. If so, they were probably using Fortran, and there's probably lots of people doing that sort of work in Fortran even now.

Not that I would recommend using Fortran for anything, just providing context. Even Numerical Recipes didn't get a C version until 1988. The original edition from 1985 was Fortran and Pascal. So even Pascal was considered a more mainstream language for this sort of work than C, at that time.

That's a good point. Even Matlab and IDL didn't have the widespread reach of Fortran for that kind of thing until later.

Fortran itself is still under development. It has updated standards from 1990, 1995, 2003, 2008, and 2018. Compilers are mostly compliant with 2003 or 2008 with 2018 features still being added. http://fortranwiki.org/fortran/show/Compiler+Support+for+Mod...

For people doing serious work in certain fields it's still a popular choice. It's still in use in industry and is a must in some academic settings. It's a bit of a niche, but then most things are.

Sure, I mean, I use SciPy, much of which is wrappers around Fortran libraries.

Reply on news.ycombinator.com