[Haskell-cafe] Haskell.org GSoC

Don Stewart dons at galois.com
Wed Feb 11 12:51:26 EST 2009


d:
> Hi,
>
> I noticed last year Haskell.org was a mentoring organization for  
> Google's Summer of Code, and I barely noticed some discussion about it  
> applying again this year :)
>
> I participated for GCC in 2008 and would like to try again this year;  
> while I'm still active for GCC and will surely stay so, I'd like to see  
> something new at least for GSoC.  And Haskell.org would surely be a  
> very, very nice organization.
>
> Since I discovered there's more than just a lot of imperative languages  
> that are nearly all the same, I love to do some programming in Prolog,  
> Scheme and of course Haskell.  However, so far this was only some toy  
> programs and nothing "really useful"; I'd like to change this (as well  
> as learning more about Haskell during the projects).
>
> Here are some ideas for developing Haskell packages (that would  
> hopefully be of general use to the community) as possible projects:
>
> - Numerics, like basic linear algebra routines, numeric integration and  
> other basic algorithms of numeric mathematics.

I think a lot of the numerics stuff is now covered by libraries (see
e.g. haskell-blas, haskell-lapack, haskell-fftw)
 
> - A basic symbolic maths package; I've no idea how far one could do this  
> as a single GSoC project, but it would surely be a very interesting  
> task.  Alternatively or in combination, one could try to use an existing  
> free CAS package as engine.

Interesting, but niche, imo.
 
> - Graphs.
>
> - Some simulation routines from physics, though I've not really an idea  
> what exactly one should implement here best.

True graphs (the data structure) are still a weak point! There's no
canonical graph library for Haskell. 
 

> - A logic programming framework.  I know there's something like that for  
> Scheme; in my experience, there are some problems best expressed  
> logically with Prolog-style backtracking/predicates and unification.  
> This could help use such formulations from inside a Haskell program.  
> This is surely also a very interesting project.

Interesting, lots of related work, hard to state the benefits to the
community though.
 
> What do you think about these ideas?  I'm pretty sure there are already  
> some of those implemented, but I also hope some would be new and really  
> of some use to the community.  Do you think something would be  
> especially nice to have and is currently missing?

Think about how many people would benefit.

For example, if all the haddocks on hackage.org were a wiki, and
interlinked, every single package author would benefit, as would all
users. 

-- Don


More information about the Haskell-Cafe mailing list