Conversation

Notices

1. Embed this notice
   screwlisp (screwtape@mastodon.sdf.org)'s status on Tuesday, 14-Jan-2025 05:39:58 JST screwlisp

   So my question is, what's an intuitive way to use dropping-in- #fortran from the #lisp #repl ?
   Reading a .f file ~ into a let* form which I'm currently doing seems kind of bland. #f( this(1) = is(2) + a / fortran * line) seems kind of uninspiring (who would want to express themselves like this). Maybe Enter "fortran mode" and read lines of fortran from *standard-input* with normal interactive evaluation hacked in?

   The fortran becomes #series expressions in lisp.

   • Embed this notice
     screwlisp (screwtape@mastodon.sdf.org)'s status on Tuesday, 14-Jan-2025 05:41:17 JST screwlisp

     in reply to

     The point is that a lot of demonstrative algorithmic programming has been done in fortran, so it's nice to be able to just drop in bits of reference implementations (of FFTs for example).

   • Embed this notice
     Alfred M. Szmidt (amszmidt@mastodon.social)'s status on Tuesday, 14-Jan-2025 05:43:48 JST Alfred M. Szmidt

     in reply to

     @screwtape How serious of an answer do you want?

     screwlisp repeated this.
   • Embed this notice
     Alfred M. Szmidt (amszmidt@mastodon.social)'s status on Tuesday, 14-Jan-2025 05:46:22 JST Alfred M. Szmidt

     in reply to

     @screwtape That was a serious question, with a hint of "this has been done before".

     (fc-file "foo.f") ;; foo.f defines FOO function.
     (foo)

     This is how Fortran, Prolog, C etc got implemented on the .... #LispMachine .. this is where it got unserious. =)

   • Embed this notice
     screwlisp (screwtape@mastodon.sdf.org)'s status on Tuesday, 14-Jan-2025 05:46:32 JST screwlisp

     in reply to

     • Alfred M. Szmidt

     @amszmidt I'm always interested in all of your opinions. I think my case of just getting Burrus' reference generated fortran prime FFTs in lisp, series optimized for what it's worth, is meaningful, so I don't think it's vacuous, but I'm happy to have your thoughts.

   • Embed this notice
     Alfred M. Szmidt (amszmidt@mastodon.social)'s status on Tuesday, 14-Jan-2025 05:49:37 JST Alfred M. Szmidt

     in reply to

     @screwtape Since you can read a Fortran file in Lisp, that already hints that things are converted to Lisp.

     #f(... fortran .. code ..) is not Lispy, it is also essentially dong double work for a LOAD like function.

     So I'd just have a LOAD (CL doesn't let you have fun .. so maybe FC-LOAD); which populates your image with the Fortran and whatever functions it defines.

   • Embed this notice
     screwlisp (screwtape@mastodon.sdf.org)'s status on Tuesday, 14-Jan-2025 05:55:38 JST screwlisp

     in reply to

     • Alfred M. Szmidt

     @amszmidt
     One particular thing I'm trying to get out of the fact that I'm also using Waters' series in lisp is that I want
     (fortran-inplace-fft *arr* :start 1 :stop 3)
     to use Series' lazy evaluation to only resolve bins 1 and 2 of the fourier transform as a sort of lazy-person's sparse transform. Though this is a separate issue, being what I'm doing with the lisp s-expressions afterward, rather than anything special about fortran code in particular.

   • Embed this notice
     Alfred M. Szmidt (amszmidt@mastodon.social)'s status on Tuesday, 14-Jan-2025 06:01:30 JST Alfred M. Szmidt

     in reply to

     @screwtape That sounds like passing Lisp data to Fortran .. that can quickly become annoying. I know this from Zeta-C...

   • Embed this notice
     screwlisp (screwtape@mastodon.sdf.org)'s status on Tuesday, 14-Jan-2025 06:35:57 JST screwlisp

     in reply to

     • DougMerritt (log😅 = 💧log😄)

     @dougmerritt
     FFI
     Didn't even try

   • Embed this notice
     DougMerritt (log😅 = 💧log😄) (dougmerritt@mathstodon.xyz)'s status on Tuesday, 14-Jan-2025 06:35:58 JST DougMerritt (log😅 = 💧log😄)

     in reply to

     @screwtape
     Link the compiled Fortran to the compiled Lisp. 🤨