Subj : Re: RIP Niklaus Wirth
To   : tenser
From : poindexter FORTRAN
Date : Fri Jan 12 2024 06:34:00

-=> tenser wrote to Dr. What <=-

 te> There's an old joke.  Q: "What language will scientists and
 te> engineers be programming in in 50 years?"  A: "I don't know,
 te> but it will be called Fortran."

 Reminds me of the XKCD comic, "Team Chat". If it works, someone will
 still use it.

 https://xkcd.com/1782/

2004:
 Our team stays in touch over IRC.

2010:
 Our team mainly uses Skype, but some of us prefer to stick to IRC.

2017:
 We've got almost everyone on Slack, but three people refuse to quit IRC
 and connect via gateway.

2051:
 All consciousnesses have merged with the galactic singularity, except
 for one guy who insists on joining through his IRC client.

 "I have it set up the way I want, Okay?!"




 DW> But other languages are trying to take over.  Julia is a good example
 DW> here.  It's very much like Python, but it's also very fast with a focus
 DW> on math.  A good potential replacement for FORTRAN.
 DW>
 DW> At some point, these applications need to be rewritten.  Oh, now I will
 DW> have nightmares about the FORTRAN IV code that I worked on in 1984 that,
 DW> even though I converted it to VS FORTRAN at that time, is still running
 DW> and hasn't been rewritten into something more modern.

 te> A friend of mine, now retired, is a former architect for
 te> high performance systems at Intel (did a PhD in transputers,
 te> worked on HPC-stuff his entire career).  He likes to say,
 te> "Fortran pays my salary."  He didn't personally program in
 te> it himself, but he had some very convincing arguments
 te> about _why_ Fortran was still used and would (and should)
 te> continue to be.  Not only is modern Fortran actually a
 te> pretty reasonable language, it turns out that a lot of the
 te> design of the language lends itself to very aggressive
 te> optimizations; the aliasing rules, for instance, mean that
 te> the compiler can automatically parallalize lots of programs
 te> in a way that other languages (so far) haven't been able to
 te> match.  Also, "the math hasn't changed."

 te> All that said, it'll be interesting to keep an eye on Julia
 te> and see where it goes.  And this is to say nothing about
 te> Matlab, Mathematica, and other interactive environments that
 te> have Fortran-like languages and are popular in science
 te> and engineering.

 te> --- Mystic BBS v1.12 A48 (Linux/64)
 te>  * Origin: Agency BBS | Dunedin, New Zealand | agency.bbs.nz (21:1/101)

.... All of my certifications are self-signed.
--- MultiMail/Win v0.52
 * Origin: realitycheckBBS.org -- information is power. (21:4/122)

.