Conversation

Notices

1. Embed this notice
   Arne Babenhauserheide (arnebab@rollenspiel.social)'s status on Sunday, 29-Dec-2024 18:53:20 JST Arne Babenhauserheide

   When you check your website logs in the morning and realize that the most read #article the day before was about #Fortran #programming:
   https://www.draketo.de/light/english/fortran-surprises

   #code #software

   • Embed this notice
     Tom Hayward (hattom@mastodon.social)'s status on Sunday, 29-Dec-2024 19:52:44 JST Tom Hayward

     in reply to

     @ArneBab lots of good stuff in there!

     "Alternative: Declare the return type of a function inside the function" -- this is what I do for any function which returns anything other than a scalar of one of the simple built-in types.

     The part about accumulation is quite subtle -- variables can have a property "save" which makes them persistent across calls.
     Any initialisation in the declaration line implicitly makes them "save", which is often counter intuitive.
     Reminds me of python's: def a(b=[]):

   • Embed this notice
     Arne Babenhauserheide (arnebab@rollenspiel.social)'s status on Sunday, 29-Dec-2024 19:52:44 JST Arne Babenhauserheide

     in reply to

     • Tom Hayward

     @hattom yes ☺

     Do you think I should extend that part?

   • Embed this notice
     Arne Babenhauserheide (arnebab@rollenspiel.social)'s status on Sunday, 29-Dec-2024 20:12:29 JST Arne Babenhauserheide

     in reply to

     • Tom Hayward

     @hattom that’s a pretty neat way to obfuscate ☺

     But I have an inkling how that may have come to be (first the explicit version, then the shorter default).

   • Embed this notice
     Tom Hayward (hattom@mastodon.social)'s status on Sunday, 29-Dec-2024 20:12:30 JST Tom Hayward

     in reply to

     @ArneBab while I'm at it with unnecessary comments, you don't *have* to set the return value of a function to a variable of the same name, but if you don't want to, you need to be explicit about what the name of the result is called.

     function myfunc result(ans)
     real :: ans
     ans = 3.0
     end function

   • Embed this notice
     Tom Hayward (hattom@mastodon.social)'s status on Sunday, 29-Dec-2024 22:14:41 JST Tom Hayward

     in reply to

     @ArneBab we have all 3 (real function X; function X + real::X; function X result(Y) + real::Y) in our production code.

     I only use the result option if the function name is quite long; or the function name and the return value name are not logical enough [e.g.: function get_integral_from_shape(shape) result(integral)].

     Definitely don't want to be writing code with lines like:
     get_integral_from_shape_cubic = get_integral_from_shape_cubic + func(get_integral_from_shape_cubic)

     Oh... I wish "+="

   • Embed this notice
     Arne Babenhauserheide (arnebab@rollenspiel.social)'s status on Sunday, 29-Dec-2024 22:14:41 JST Arne Babenhauserheide

     in reply to

     • Tom Hayward

     @hattom yes — do you know why that’s not included? Will it come in Fortran 20xy ?

     Besides: I have a short article that pulls my three main texts on #Fortran together:
     https://www.draketo.de/software/learn-fortran

     I thought you might like it ☺

   • Embed this notice
     Arne Babenhauserheide (arnebab@rollenspiel.social)'s status on Sunday, 29-Dec-2024 22:31:25 JST Arne Babenhauserheide

     in reply to

     • Tom Hayward
     • Jeff

     @hattom not pronounceable sounds like limiting through English: German does that easily: "plusgleich" ☺

     Thank you for the info!

     @jeffscience

   • Embed this notice
     Tom Hayward (hattom@mastodon.social)'s status on Sunday, 29-Dec-2024 22:31:26 JST Tom Hayward

     in reply to

     • Jeff

     @ArneBab iirc, @jeffscience wrote something about it before -- committee argued against it for some strange reason (because it's not pronounceable??) and shot down the proposal.

     Thanks for the link, I'll take a look!

   • Embed this notice
     Arne Babenhauserheide (arnebab@rollenspiel.social)'s status on Sunday, 29-Dec-2024 23:35:27 JST Arne Babenhauserheide

     in reply to

     • Tom Hayward

     @hattom for a formula translation, that makes sense, yes …

   • Embed this notice
     Tom Hayward (hattom@mastodon.social)'s status on Sunday, 29-Dec-2024 23:35:28 JST Tom Hayward

     in reply to

     • Jeff

     @ArneBab @jeffscience I really don't remember for sure. But I think it wasn't quite that, but rather that it's not a normal mathematical operation; traditional operators have been words. I don't know where I read the discussion, but it was a long time ago.