Re: [eigen] Eigen2 in Avogadro

["Gael Guennebaud" <gael.guennebaud@xxxxxxxxx>]

On Mon, Aug 25, 2008 at 4:32 PM,  <jacob@xxxxxxxxxxxxxxx> wrote:
> ouch! You checked out all of it? There is also lxr.kde.org but indeed it is
> less powerful i.e. it seems to search in trunk+branches and there is
> apparently no way to restrict to trunk.

actually, once the source are in the cache it is only a matter of a
few seconds...

>> So what next ?
>
> For me, I tell you frankly: I've delayed my focusing-on-job too much
> already, I cannot continue coding for now besides kritaplugins/panorama.
> I'll try to contribute to Eigen's documentation (btw kudos for the
> QuickStartGuide and the rest, it's very good). I can't tell exactly how bad
> this is going to be.

same here, I really really should calm down Eigen hacking ! I know I
said that for a while... but I had hard time to get focused on my job.
For the next 3 months I'll do only documentation and bug fixing too.

> About kgllib: sure, it is a very good idea to port it to eigen2. According
> to lxr it is currently only used in Astrododge, a new game by Rivo currently
> in playground, but as with other libs this is a chicken and egg problem and
> so it's definitilely worth porting it (sidenote: kgllib reuses some code
> from avogadro.)

I contacted Rivo, let's see.... (forgot to cc the list but I'll keep
you informed !)

>> step uses gmm
>
> This one is a top priority because it is part of the kde base modules
> (namely kdeedu). And also because it's a cool program and because its
> authors already expressed interest in Eigen. But, it would use the Sparse
> module. At the KDE-Edu meeting in december I talked with the Step authors
> (the Kuznetsov) and this is what told me they were using:
> -gmm::cg, constrained_cg
> -dantzig algorithm
> -types: gmm::row_matrix, col_matrix, rsvector, array1D_reference

good to know. Daniel (aka cylmor), are you still interested by hacking
the Sparse module ? this might gives you a hint on what to do first !
actually our dense and sparse matrices already fulfill the storage
needs. Beyond polishing the Sparse API, what's really missing are the
high level solvers...

>> playground/solidcreator uses eigen1
>
> Hehe, this one is by Casper Boemann, one of the co-founders of Eigen. He
> already asked me if I could port it to eigen2, but I had no time. If you do
> it he would be very happy.

unfortunately I'm not going to be able to do it soon  :(  I've already
enough to do !

gael.