[SciPy-User] Least-squares fittings with bounds: why is scipy not up to the task?

[Ralf Gommers]

On Sat, Mar 10, 2012 at 3:01 PM, Nathaniel Smith <njs at pobox.com> wrote:

> On Fri, Mar 9, 2012 at 9:36 PM, Ralf Gommers
> <ralf.gommers at googlemail.com> wrote:
> > I don't think anyone has doubts about the quality of lmfit. On the
> contrary,
> > I've asked you to list it on http://scipy.org/Topical_Software (which
> you
> > did) because I thought it looked interesting, and have directed some
> users
> > towards your package. The documentation is excellent, certainly better
> than
> > that of many parts of scipy. The worry with your code is that the
> > maintenance burden may be relatively high, simply because very few
> > developers are familiar with AST. The same for merging it in scipy - one
> of
> > the core developers will have to invest a significant amount of time
> > wrapping his head around your work.
>
> Out of curiosity (and apropos an earlier thread), would it affect your
> reservations if lmfit's ad-hoc AST usage and python interpreter were
> replaced by a simple call to 'eval'?


I don't think using "eval" will make anyone happy.

Also, my reservation isn't very strong. I personally think it would make
sense for lmfit to remain a stand-alone package for a while, but if others
disagree and there's one developer who invests the required time in
reviewing/merging it, that's all it would take. This doesn't even have to
be a core scipy developer - if you for example would do this work and
indicate that you would be able to do some basic maintenance for it in case
Matt can't / won't anymore, that would be perfectly fine.

Cheers,
Ralf
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.scipy.org/pipermail/scipy-user/attachments/20120310/7c0c31c9/attachment.html>