We have a GSoC mentor already: David Simcha

[dsimcha]

On 3/2/2012 9:38 PM, Trass3r wrote:
> Am 03.03.2012, 00:43 Uhr, schrieb Andrei Alexandrescu
> <SeeWebsiteForEmail at erdani.org>:
>
>> David Simcha applied for a second gig as a GSoC mentor. Needless to
>> say, his application was approved :o). Please join me in welcoming him!
>
> Yay!
>
> Time to ask about the status of the last GSoC project, i.e. the LinAlg one.
> If it still needs lots of work, maybe there could be another round on that.

The status is that the debugging and polishing is slowly happening. 
I've used the library for real work and while rough around the edges, 
it's quite good.  I've also worked on adding some Lapack wrapper stuff 
that Cristi (the student I mentored) didn't get to and improving the 
test suite.

The key todos are:

1.  Fix a few nasty bugs that are more design flaws than run-of-the-mill 
bugs.  This is hard for me to do without Cristi's input because I am 
unclear on a few design decisions.

2. Documentation.

3  More LAPACK wrappers.

4.  More real-world testing.  I'm not comfortable submitting something 
this large and complicated for Phobos inclusion until a few people have 
used it extensively for real work and found all the bugs and design flaws.

5.  Some serious profiling/performance optimization.

6.  Get allocators into Phobos, since Cristi's SciD fork depends on them.