[SciPy-dev] Re: [Numpy-discussion] Current thoughts on future directions

Travis Oliphant oliphant at ee.byu.edu
Thu Mar 10 11:50:39 CST 2005

Stephen Walton wrote:

> Can I put in a good word for Fortran?  Not the language itself, but 
> the available packages for it.  I've always thought that one of the 
> really good things about Scipy was the effort put into getting all 
> those powerful, well tested, robust Fortran routines from Netlib 
> inside Scipy.  Without them, it seems to me that folks who just 
> install the new scipy_base are going to re-invent a lot of wheels.

I didn't emphasize it, but of course there will still be the rest of 
scipy available wrapping all the fortran stuff (and more).  This will be 
advertised at the site where people download scipy_core.   If they don't 
need it they won't download it, but they will be aware of its existence. 
  It will just be a separate package (or set of packages).  In other 
words we go from today's Numeric + scipy  to tomorrows scipy_core + 
scipy.  So, this really isn't different from the current experience for 
scipy users.

> Is it really that hard to install g77 on non-Linux platforms?

I think what Konrad (and I know of at least 3 others) are saying is that 
it is hard enough that he doesn't want to require it for users of his 
packages, especially when his package doesn't need it.   I think the 
argument has merit.  

The current strategy doesn't really change anything for current scipy 
users.  But, it does make it possible for more people to start using 
scipy_core and have a common package that nearly all of the community 
can get behind.


More information about the Scipy-dev mailing list