Hi all,
This is Don, the Microsoft Research author of F#. It's early days yet for F# so I'm glad to hear what people have to say. F# is experimental around the edges, but I wanted a solid, sensible core that no one needed to argue about.... [ Actually, to be really honest, I just wanted a.NET compiler that meant I could keep programming the way I wanted to program and could reuse my existing investment in code, so thought "what the hell, I'll write a compiler..." ]
For a version of ML (in this case Standard ML) with all the features such as functors enabled by using whole-program compilation please take a look at SML.NET - email Andrew Kennedy (akenn@microsoft.com) for more details. Andrew and co. will be posting a release in the next week or so. It rocks!
OCaml is a moving target, as they are adding new features quite often, and some of the features I left out of F# were because I don't actually like them very much (labels for example - in this case something may be needed, but I don't think they got it quite right, at least not in the context of accessing.NET libraries.) But F# has all the features of OCaml that I actually like to use, with the exception of functors which I occasionally use very sparingly (and some day I'll probably add the equivalent functionality to F#).
I think the first poster made an excellent point: simpler languages can work well in some areas of the modern programming landscape: adding new features to languages can give diminishing returns. Us programming languages researchers have been very guilty of forgetting how well tools such as Perl (and countless other scripting languages and small, special-purpose languages) have served their users.
My apologies if I can't track all of this discussion,
Don
Slashdot Mirror
Oh yuck, I'd better practice my HTML.
Hi all, This is Don, the Microsoft Research author of F#. It's early days yet for F# so I'm glad to hear what people have to say. F# is experimental around the edges, but I wanted a solid, sensible core that no one needed to argue about.... [ Actually, to be really honest, I just wanted a .NET compiler that meant I could keep programming the way I wanted to program and could reuse my existing investment in code, so thought "what the hell, I'll write a compiler..." ]
For a version of ML (in this case Standard ML) with all the features such as functors enabled by using whole-program compilation please take a look at SML.NET - email Andrew Kennedy (akenn@microsoft.com) for more details. Andrew and co. will be posting a release in the next week or so. It rocks!
OCaml is a moving target, as they are adding new features quite often, and some of the features I left out of F# were because I don't actually like them very much (labels for example - in this case something may be needed, but I don't think they got it quite right, at least not in the context of accessing .NET libraries.) But F# has all the features of OCaml that I actually like to use, with the exception of functors which I occasionally use very sparingly (and some day I'll probably add the equivalent functionality to F#).
I think the first poster made an excellent point: simpler languages can work well in some areas of the modern programming landscape: adding new features to languages can give diminishing returns. Us programming languages researchers have been very guilty of forgetting how well tools such as Perl (and countless other scripting languages and small, special-purpose languages) have served their users.
My apologies if I can't track all of this discussion,
Don