> To the newcomer who is not part of the FP academic community, this all
> makes life sort of difficult. These differences seem larger than the
> differences among C compilers and are MUCH larger than the differences
> among Java compilers. I have been trying to learn Haskell and have been
> impressed with both its elegance and the way it allows me to write code
> that works on the first try (or two).
These are fair comments. In my view the things that are holding back
Haskell are not language issues at all. They are
- foreign language interfaces
- libraries
- programming environments
There are a bunch of people working on all of these, but the results
are not yet stable. In particular
* On foreign language interfaces, we have now been through several iterations
of Green Card, and are poised for what I hope is our final "take",
currently called H/Direct. There's a paper on my Web page, and an
implementation on the way. http://www.dcs.gla.ac.uk/~simonpj
[I say "we" meaning "Glasgow plus many
colleagues", notably Yale, York, Utrecht.]
Foreign language interface implementations have certainly been a
moving target. It would be better to get it "right" first time,
but worse to have "standardised" prematurely.
* On libraries, there is quite a lot going on. GHC and Hugs have some
more or less standardised libraries going well beyond the Haskell 1.4
set. The Binary library is a hot topic. E.g. Glasgow and York had
a full-day meeting on this very topic two weeks ago; that should lead
to a library proposal (based on the current York story)
http://www.cs.york.ac.uk/fp/nhc13/binary.html
for the world to discuss/critique, and thence to implementations
for GHC and Hugs as well as NHC.
As ever, these things take time.
* On programming environments, we are working hard on integrating GHC
and Hugs. It's coming, promise.
Of course, all this just describes stuff I know about. There are probably
other people beavering away on good stuff; please sing out!
> However, I am not a researcher. I do commercial software development and
> need some documentation and stability. I understand that everything here
> is fairly new, but it would be really usefl if someone would post a
> summary of the FP community politics for those of us outside the research
> community:
> * What verions of Haskell should someone outside the research community be
> using. e.g. is GHC ver x particularly experimental? Is Haskell as a
> whole really a research language and not ready for commercial
> (e.g. high intensity server based) applications?
Hmm. GHC 2.10 is rather solid. GHC 3.01 is more experimental.
http://www.dcs.gla.ac.uk/fp/software/ghc/
I would love for Haskell to be used for high-intensity applications,
but before betting my company on it I'd protype my application.
Our research priorities these days are heavily influenced by
what people say they need; so talk to us.
> * Any particular language features that we should avoid or that are
> likely to change?
In Haskell 1.4 the main proposed simplification is that [x | x<-xs]
will mean a *list* comprehension rather than a *monad* comprehension.
I think the libraries are much more volatile than the language.
> * Who "owns" haskell? How are decisions made about what goes in/out of
> the language? Who controls the content of Haskell.org?
The whole thing is pretty informal. There no "official" standardisation
body etc. Instead, there is a working group chaired by John Hughes,
which is trying to develop "Standard Haskell". More info at
http://www.cs.chalmers.se/~rjmh/Haskell
The discussion is readable by all, but only writable by the working
group. This is admittedly only semi-democratic; but it is a way
of making progress.
As Will says, I think the whole scene is pretty exciting. The Haskell community
works together well; and the focus is increasingly on making a *usable*
language. The main difficulty is simple lack of available effort: anyone who
wants to contribute to any of these things, speak up!
All of this reflects my own personal views; I don't claim to speak
for anyone else.
I hope this helps.
Simon
