[SciPy-user] Arrayfns in numpy?
Tue Mar 6 13:37:31 CST 2007
This is not about adding new functionality to numpy. It's about keeping
something that was in Numeric (arrayfns.interp, in particular), and that was
taken out for numpy.
Whereas porting code to numpy for Numeric is trivial
in most cases, if the code included arrayfns.interp then a dependency on scipy
must be added (which is a big deal!), or a new interpolation function must
be written from scratch.
To make it obvious that the new numpy should not
include computations, perhaps this functionality should sit in the oldnumeric
compatibility layer. Would that be satisfactory to Souheil?
--- SciPy Users List <firstname.lastname@example.org wrote:
> I am strongly
opposed to mixing computations into NumPY. Why is 1-d
good to have inside of NumPy? why not nd-interpolation?
> why not fft?
or SVD? or some linear algebra thing?
> Python is good because of it
lets you have a lot of granularity.
> NumPy should stick to what it's
good at. And it should be very good
> at a small number of things, namely
it should be a library for
> efficient storage and access to data that
is well represented by N-
> dimensional arrays.
> Souheil Inati, PhD
> Assistant Professor
> Center for Neural Science
> New York University
> 4 Washington Place, Room 809
> New York, N.Y.,
> Office: (212) 998-3741
> Email: email@example.com
> On Mar 6, 2007, at 1:02 PM, Glen W. Mabey wrote:
> > On Tue, Mar 06, 2007 at 09:43:27AM -0700, Travis Oliphant wrote:
> >> Perhaps we could put interp in there as well (it doesn't look too
> >> big).
> >> A simple 1-d interpolation would probably be a useful
> >> have in
> >> NumPy. What do others think?
> > +1
> > _______________________________________________
> > SciPy-user
> > SciPyfirstname.lastname@example.org
> > http://projects.scipy.org/mailman/listinfo/scipy-user
> SciPy-user mailing
More information about the SciPy-user