Hacker Read top | best | new | newcomments | leaders | about | bookmarklet login
How do Scala and F# compare when it comes to application development features? (www.developer.com) similar stories update story
7.0 points by gkunene | karma 43 | avg karma 1.43 2010-07-01 20:21:27+00:00 | hide | past | favorite | 5 comments



view as:

These types of comparisons seem to be better for checkbox style decision-making. It's something I can put onto a spreadsheet for a manager whose decision-making process will be "Scala has 5 check marks, F# has 6. F# it is!"

Good read! I've been meaning to give F# a try were it not for the other projects I've been juggling in my spare time. Have any HN developers taken it out for a spin?

I've only used F#; have looked at Scala a bit but haven't used it. I find F# feels fairly natural and pleasant if you know C# and like ML-like languages. I've mostly treated it as a nicer C# rather than a true functional language, though; features like sequence expressions (generate an anonymous IEnumerable with yields without having to declare and define a function etc.) and object expressions (generate an instance of an anonymous class that inherits from some interface) let you write imperative code in a more "value-oriented" fashion and with less ceremony. Algebraic data types and pattern-matching are also nice to have, of course. My impression is that Scala is much more conceptually interesting.

Meh. I'm not impressed with either this article or the preceding one.

I've been programming almost exclusively in F# over the last year - but they've been smallish projects. The weird thing about F# is that the code keeps collapsing from huge constructs down to really small pieces.

A couple of nits from both of this guy's articles. F# supports multiple inheritance through interfaces, which ain't C++ but it's good enough for government work. F#'s pattern-matching is freaking way far above the simple case he showed there: active patterns, assignment on test, structural-matching, etc. You can do some super cool stuff with F#'s pattern-matching. Hopefully he knew this and it was just an oversight.

Also he seems to indicate that MS is going to use F# as an entrance into large-scale scientific-type computing. This may be true, but it is badly misleading. By teaching developers to work with immutable data and functional composition, it opens up the entire PC stack to all sorts of improvements over your average VB code. I expect to see some really neat stuff come out along these lines.


Legal | privacy