Re: [Gretl-devel] arch as summer of code project?
On Fri, 12 Mar 2010, Sven Schreiber wrote:
Ok thanks for the information. I looked at the SoC FAQ and I now
have
the impression that it's quite "organization-centric" as opposed to
"project-centric". So I don't think it will make much sense to pitch
this Arch thing. They want a pool of ideas (which you provided last
year), and I think just applying again with a basically unchanged ideas
pool is not going to work. I think a viable plan could be to embed the
SoC application (next year) in the "big gretl release" thing. Then there
would almost automatically be a large pool of ideas. It could also be
combined with the next gretl conference. However, that goes against
Jack's timeline since that would mean to wait one and a half years. (And
I also think that would be too late.)
I don't consider work on garch models a matter of life and death. It's
just something that I consider indispensable to call gretl "100% feature
complete" and that I'd like see done in 2010.
GSoC would be nice for this in theory, but in practice I consider it
unlikely that someone totally unfamiliar with gretl's codebase could
implement a whole array of estimation methods with a few months' work,
unless we're talking of a true coding genius (and IMO the chances of
involving Linus Torvalds or Fabrice Bellard are slim).
A worthwhile intermediate project for models of conditional
heteroskedasticity (much more feasible in a limited timeframe) may be a
good shakeup of Stefano Balietti's gig code, which is all in gretl
scripting language and therefore much more accessible than raw C. If
someone decided to pick that up, I'd be very very happy to act as
mentor/collaborator.
Riccardo (Jack) Lucchetti
Dipartimento di Economia
Università Politecnica delle Marche
r.lucchetti(a)univpm.it
http://www.econ.univpm.it/lucchetti