On 10/29/11 10:24 AM, dsimcha wrote:
On 10/29/2011 10:59 AM, Andrei Alexandrescu wrote:
I understand. The way I see it, if SciD can't be considered functional
without triangular and symmetric matrices, then we counted an unfinished
project as successfully completed. This is an imperfection of our GSoC
planning and scrutiny process that we should fix in the future.
We shouldn't expect and wait that Cristi will restart work on the
project. His future participation would be of course gladly welcome, but
we should make nothing contingent upon that.
With that in mind, let me ask again: is there someone in the community
willing to do what it takes to put SciD through the Phobos review
process?
Thanks,
Andrei
If Cristi never gets around to resuming work on SciD, I'll work on it
slowly but surely. I agree SciD is unfinished, but this was at least
partly due to unforeseen things like compiler bugs and the last-minute
decision to change things over to COW/value semantics and other misc.
scope creep. IMHO Cristi did a very good job under the circumstances and
there's no way I could complain/fail him for not finishing
triangular/symmetric matrices before the GSoC deadline, if that's what
you're implying. Writing a full scientific lib in a summer is pretty
ambitious and I'm just glad that we got so far off the ground.
Oh, no question about it. Cristi has done a great job. We all did,
considering the lack of experience and the constraints. Ideally, we
would have figured the best corrective actions and features to cut to
materialize the impact of the project by the end of summer. That impact
has for now remained in a latent form that needs to be realized.
One great thing you could do at this point is to sketch a roadmap of the
project from where it is right now to a rounded, deliverable library.
That way anyone interested would know what steps to take.
Thanks,
Andrei
