Back to Questions
1.9k
views10
comments
VS2010 Support for F#
f#

I've been pretty disappointed with the support for F# in VS2010 - to the point that I wonder if there are some capabilties that I don't have. In particular, I was surprised to find (or more specifically, not find) things that you take for granted in C# like outlining, find all references, rename support, navigation via dropdowns that index the module's content, etc.

A particular example that causes me problems is the use of dependency injection which is particularly natural in an F# environment with first class functions. Unfortunately, the only useful mechanism I'm aware of for finding a function definition is 'Go to definition' which of course stops at the point of refererence in DI leaving you to remember which module the actual function might be in and worse, text search to find the function in that module.

I realize that many features may have been lost in the push to get F# support into VS, but it seems to me that F# is such a 2nd class citizen in VS that it's benefits relative to C# are sometimes hard to discern.

Is it possible that I am missing features that actually exist? Is there a plan to make better support available in the near future?

--Don

.

Well, then it's not so different from F# for VS 2008.
Why then F# for VS 2010 was announced as first class language in VS 2010 ???

By on 3/12/2010 5:01 AM ()

Damn, I was looking forward to VS2010 final (haven't tried the betas/RC) for a better support to F#!
:-(

By on 3/12/2010 5:24 AM ()

The F# visual studio integration is built as a plugin, so exactly the same code runs in Visual Studio 2008 as 2010. So if you've tried the latest builds of the F# plugin, you're already using the features that will be present in VS2010. The only real advantage of VS2010 for an programmer is that you can use F# with the .NET framework 4.0.

The F# integration we have today is considerably better than when the productization effort began, a little under 2 years ago (if I remember correctly).

It's true that the F# integration is not as feature rich as the C# integration, but the F# team aren't doing so badly when you consider that the C# team have had a 5 year head start on them.

I also disagree that F# has few advantages as programming language. F# has pretty much all the feature of C# (the only real exception I can think of is anon. types) and many features that C# doesn't have. To name but a few of these:

- pattern matching,

- union types,

- record types with copy and update semantics,

- object expressions for creating objects in line

I think it's being called a 'first class' visual studio language simply because its available by default.

Cheers,

Rob

By on 3/12/2010 7:39 AM ()

There's one sure-fire cure for any lack of F# support in future versions of Visual Studio:

We, the current and learning F# programmers, need to make the language such a critical part of the programming landscape that lack of support is unthinkable. To that end, we need to support the use and teaching of the language in business, programming education, research and in the online community.

Microsoft has done a good job at getting the ball rolling; now it's in our court.

-Neil

By on 3/12/2010 9:41 AM ()

I'd say more than anything the lack of partial classes on F#'s part and the lack of language-agnostic code generation in VS' design tools makes it virtually impossible for F# to ever become a first class citizen.

I've brought this up ad nauseam, but I really think partial class support, nested classes and more flexible mutual dependencies need to be top priorities for F#.

By the way, I think that certain things -- for instance the fact that many if not most of the operators are defined as part of the libraries rather than syntax -- make things like IntelliSense difficult.

By on 3/13/2010 4:24 PM ()

Well, may be F#'s implementation and its syntax make its integration in VS 2010 so hard.
Does Microsoft intend to develop F# further, or it is just a research project???

By on 3/14/2010 6:56 AM ()

Well, may be F#'s implementation and its syntax make its integration in VS 2010 so hard.
Does Microsoft intend to develop F# further, or it is just a research project???

Perhaps F#'s syntax does make integration into visual studio harder, especially for things like autocompletion, go to definition etc. However with the VS 2010 release they have achieved a good level of integration into visual studio, even if it is not as fully featured as the C# support.

It is also now a product supported by Microsoft, so I think it safe to say it has moved beyond a research phase.

I believe the F# team have already started working on the VS next release. I believe they intend to develop both new language and Visual Studio integration features. So, yes Microsoft are planning to continue developing it.

Thanks,

Rob

By on 3/15/2010 8:31 AM ()

I certainly hope Microsoft continues to mainstream F#. This is a complete reversal from how I felt several months ago when I first started exploring F#.

There's no polite way to put this, but the first encounter with F# is a real learning cliff. This may turn back a lot of potential users who browse the language casually.

However, once one gets over that cliff, F# is a wonderful language. I'm finding I would now much rather program many things in F# than in C# or C++. And I think in the long run, it will be much easier for me to "think" in F# even when I'm exploring algorithmic and design problems away from the computer.

But the key is to find ways to get people over that first cliff. That will generate the demand Microsoft needs to justify keeping F# mainstream.

-Neil

p.s. If you're new to F#, Chris Smith's [i]Programming F#[/i] (O'Reilly, 2010) is a great resource. It made a big difference for me.

p.p.s While we're on the wish list, here's something I'd like (though I won't be too disappointed if it doesn't happen):

Arbitrary operator definition with control over arity and associativity, as is found in some versions of Prolog.

By on 3/14/2010 8:05 AM ()

I always liked functional programming. Now, F# is a real opportunity to apply functional style of programming in real world applications. But the question is, will Microsoft support the language in the future? Is the language e long term project, or just experimental one? Does the effort to learn it and program in it make sense?

By on 3/14/2010 11:14 AM ()

Does the effort to learn it and program in it make sense?

When I learned Haskell in 1997, I couldn't use it in real world projects, but it really improved my C++ coding style, especially when using STL. So, I'd say it's definitely worth the effort to learn one functional, one logical and one dynamic language (did I miss anything?) in order to broaden your perspective on programming.

By on 3/15/2010 7:10 AM ()

    Topic tags

    IntelliFactory Offices Copyright (c) 2011-2012 IntelliFactory. All rights reserved.
    Home | Products | Consulting | Trainings | Blogs | Jobs | Contact Us | Terms of Use | Privacy Policy | Cookie Policy     		Built with WebSharper