[Haskell-cafe] Re: Why Haskell?

[Bulat Ziganshin]

Hello Johan,

Monday, August 7, 2006, 7:25:47 PM, you wrote:
(sorry for too late answer)

>> http, smtp and other networking protocols - yes. xml/sql is too large
>> things. actually, haxml package is already included in GHC sources
>> distribution and i think that it should be excluded from there because
>> it's too large, far more than any other package bundled with GHC

> The problem I'm having with SQL right now is that there are a number
> of not complete and splintered implementation efforts. Having one
> library outside GHCs libraries but still promoted as the default
> implementation (and hosted under haskell.org) would be enough.

that is because there many developers what had different vision. you
can't force them to work together. all that we can do is to declare
one library as "standard" one so that the developers who want to
contribute to some DB library will work on improving this concrete lib

moreover, competition is the only source of progress, so having
several competing libs might be better actually than having one
"ultimate" one

> Same
> for XML. I think most of my problems would be solved if cabal-get
> turned out to something like Ruby Gems so I could easily pull the
> packages I need to the machines I need (and also get dependencies and
> a way to remove the packages without leaving files scattered on my
> file system).

one of Google SoC projects is to implement this cabal-get at last end.
i hope that situation will change during a next month (when a summer
will end :) )

returning to the question discussed by me - i'm all for having better
libs and make some of them "standard". but i'm against distributing
with GHC any libs that is larger than base lib. to be exact, i want
that all multimedia/gui/xml libs will be excluded from ghc, including
opengl-based, fgl, openal, object-io, haxml


-- 
Best regards,
 Bulat                            mailto:Bulat.Ziganshin at gmail.com