Jim Burton wrote:
Andrew Coppin wrote:
The other downside is that you end up with a world where most of the "tools" are in fact one-man research projects or small toys.
There are a few good, powerful, useful things out there. (GHC and Parsec immediately spring to mind.) But there's also a vast number of really tiny projects which don't seem to be terrifically well supported. Kind of makes me sad; Haskell seems almost doomed to be a language with fantastic potential, but little real-world traction.
AFAIK Haskell wasn't designed for real-world traction in the first place, but as a way of consolidating FP research efforts onto one platform, so in that sense it's a resounding success rather than "doomed". It also seems to have gained some traction, and we know that FP can be an eminently practical real-world secret weapon, so the tools you're waiting for someone else to write could well be on their way. At the same time, the only evidence for this at the moment is a lot of blogs, O'Reilly investing in a book and Eternal September on haskell-cafe. If you want a language with a bigger user base or that is less confusing, there are plenty to choose from.
The *language* I love. Haskell is usually a joy to program with. The lack of real-world traction can be very frustrating though. It's like I just found the perfect programming language, and I can't really use it for very much... :-(