[SciPy-dev] an atttempt at sandbox/xplt

Travis Oliphant oliphant.travis at ieee.org
Thu Feb 2 14:33:14 CST 2006


Robert Kern wrote:

>Nils Wagner wrote:
>
>  
>
>I'm sure there are reasons to want to move xplt back into scipy/Lib/. However, I
>think there are better reasons for keeping it out. 
>
I think the scipy library is best thought of as a toolbox of calculation 
routines.   Some of the commands may produce plots as output, but the 
user will have to supply some plotting library that just conforms to a 
(yet-to-be?) established interface.

While originally it was envisioned that something like matplotlib would 
grow inside of scipy itself, that's not what happened (and we are 
probably all better for it).

However, xplt still fills a need not met yet by matplotlib (it's *very* 
fast comparitively and can do simple 3d surface plots). Thus it should 
have a home somewhere.  I'm way too in love with SVN to use sourceforge 
for its home until sourceforge gets their SVN program out of beta.   It 
should also be recognized that xplt is pygist + some fixes + python 
tools so in an ideal world pygist and xplt would just merge. 

I should clarify a mistaken notion that Robert has though.  Despite its 
name, xplt *is cross-platform*.  It works on Windows as well as X.  
There is partial support for Mac-OSX, as well.   So, while I've never 
been in-love with xplt, it can still have a useful role to play.   I 
actually don't mind if it stays in the sandbox, but I think releases 
should be made separately and not under the scipy namespace.

-Travis





More information about the Scipy-dev mailing list