These are chat archives for Microsoft/visualfsharp

3rd
Mar 2016
Sergey Kostrukov
@kostrse
Mar 03 2016 06:20
@vasily-kirichenko DIdn't get you guys, what are the reasons anyway? :)
Vasily Kirichenko
@vasily-kirichenko
Mar 03 2016 06:45
  1. The compiler codebase is awfully complicated, only Don can make big changes in it. 2. Don thinks that the language is in a good shape and things like HKT or macros or <put anything here> should not be incorporated.
Sergey Kostrukov
@kostrse
Mar 03 2016 07:43
@vasily-kirichenko makes sense, thanks!
Vasily Kirichenko
@vasily-kirichenko
Mar 03 2016 07:45
I've been with F# for more than 3 years and I'm looking for greener fields now because I'd like to use an evolving language, not a finished one.
Sergey Kostrukov
@kostrse
Mar 03 2016 07:46
what are the prospective candidates? :) Clojure?
Vasily Kirichenko
@vasily-kirichenko
Mar 03 2016 07:47
Nemerle
Sergey Kostrukov
@kostrse
Mar 03 2016 07:49
Is it still alive?
Vasily Kirichenko
@vasily-kirichenko
Mar 03 2016 07:49
Yes, they are writing Nitra - a language workbench in which they will reimplement Nemerle from scratch.
It will allow practically limitless language extension.
That's why I'm interested in it :)
Sergey Kostrukov
@kostrse
Mar 03 2016 07:51
Nice to know, thanks, will check it out. I thought that it was dead for long after it lost interest from RSDN :))
Ah, I see JetBrains took over it... ok thanks for the information
Vasily Kirichenko
@vasily-kirichenko
Mar 03 2016 07:52
They have been working on Nitra for about 2 years at JetBrains. JB has abandoned the project a couple of weeks ago, so it's now independent OSS project.
They will start to implement Nemerle and extendable C# soon. I think we will see results in 6 months or so.
So, it's not ready to look at yet :)
It will include VS support as well.
Steffen Forkmann
@forki
Mar 03 2016 11:24
and will bring world peace.
Vasily Kirichenko
@vasily-kirichenko
Mar 03 2016 11:32
it will bring very powerful metaprogramming on .net. That's all.
Ivo Graber
@3dGrabber
Mar 03 2016 15:27
  1. The compiler codebase is awfully complicated, only Don can make big changes in it.
This sounds very unfortunate...
Is this a reason why development of the refactoring tools seems to have stalled? (Lack of open compiler a la Roslyn)
The one thing that holds me back of using F# for serious work is lack of tooling (ReSharper)
Jared Hester
@cloudRoutine
Mar 03 2016 15:31
VFPT is plenty, reshaper is needed because C# is full of boilerplate and cruft
sremani
@sremani
Mar 03 2016 16:27
rookie question, why is HKT not implemented in F# ? Is it because of CLR types ?
Loïc Denuzière
@Tarmil
Mar 03 2016 16:35
that's the main reason, yes
Jared Hester
@cloudRoutine
Mar 03 2016 16:41
I've gotten the impression Don thinks they would overcomplicate the language and make it too inaccessible. They can (and have been) been implemented as erased types in F#
Vasily Kirichenko
@vasily-kirichenko
Mar 03 2016 16:44
Personally, I don't miss HKT. Maybe because I've never seriously used a language with them.
Macros, on the other hand, would open a door for adding features to the language. It's even more important considering how hard it to do it currently, right in the compiler codebase. But Don does not hear us :(
Jared Hester
@cloudRoutine
Mar 03 2016 17:02
i'd be happy with just some simple marcos for CTFE metaprogramming type generation that gets checked properly, it doesn't even need to be flexible enough to define all kinds of new syntax constructs and whatnot
Vasily Kirichenko
@vasily-kirichenko
Mar 03 2016 17:13
Yes. F# is light years behind macros which would able to define new syntax.
Jared Hester
@cloudRoutine
Mar 03 2016 17:18
there could be a bit more freedom within the context of cexprs though
Vasily Kirichenko
@vasily-kirichenko
Mar 03 2016 17:22
Yes, but it's not enough. Although Nemerle has extraordinary metaprogramming abilities, the codebase is awful, so they saw no way to maintain and extend it further. Hence rewrite. It's sad to say, but I think the current F# compiler has no future.
Same thing has happened to C#.
It's interesting that Don does not see it.
Jared Hester
@cloudRoutine
Mar 03 2016 17:32
it seems like it's be better to do extend the language in that kind of manner once there's a stronger foundation of platform stability
Vasily Kirichenko
@vasily-kirichenko
Mar 03 2016 17:46
never can help with rotten codebase, only rewrite.
I'm not negative, I've just watched what's been happening to F#-the-language for 3 years. Nothing.
Rodrigo Vidal
@rodrigovidal
Mar 03 2016 17:52
@vasily-kirichenko I agree, in fact type providers were very disappointing to me.
Vasily Kirichenko
@vasily-kirichenko
Mar 03 2016 18:06
@rodrigovidal I don't understand F# compiler codebase very well, but it seems that TPs have polluted everything :(
It's especially sad because TPs are so limited, hard to read and do not evolve.