[SciPy-Dev] Design of scipy.stats

[Ralf Gommers]

On Tue, May 25, 2010 at 8:13 AM, David Goldsmith <d.l.goldsmith at gmail.com>wrote:

> On Mon, May 24, 2010 at 4:58 PM, David Goldsmith <d.l.goldsmith at gmail.com>wrote:
>
>> OK, having looked at the source code (duh), I have a better idea what's
>> going on; correct me if I'm wrong: the idea was to make each distribution
>> generic up to its specification parameters, and immediately available upon
>> importing scipy.stats (i.e., no need to explicitly import distributions, no
>> need for the user to create an instance of the desired distribution).  The
>> former suggested the object model (i.e., classes) be used, the latter
>> required such classes to be instantiated at module scope on import of
>> distributions (which import scipy.stats does).  With this in mind, I'm
>> comfortable w/ the status quo as far as the code is concerned, supplemented
>> w/ an explanation of this (which I'll write, unless someone feels I won't do
>> it justice - don't be shy about telling me so, I won't take it personally)
>> in (module) distribution's docstring.  I do feel, however, that it would be
>> best if the Wiki didn't think that the instances required/possessed
>> docstrings, or at least warned unsuspecting editors that editing those
>> objects' docstrings is a waste of time; accordingly, I'll be filing a
>> ticket...
>>
>> DG
>
>
> Duh (again) I'll just change the status of each instance object to
> Unimportant.
>
> That doesn't fix everything, the x_gen classes should also not be edited
right now. Plus they're not really unimportant. A new way to 'lock' pages
would really be helpful.

Ralf
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/scipy-dev/attachments/20100525/6a9644bb/attachment.html>