[Numpy-discussion] Latest Array-Interface PEP

Charles R Harris charlesr.harris at gmail.com
Fri Jan 5 10:27:45 CST 2007


On 1/5/07, Travis Oliphant <oliphant at ee.byu.edu> wrote:
>
> Charles R Harris wrote:
>
> >
> >
> > On 1/5/07, *Stefan van der Walt* <stefan at sun.ac.za
> > <mailto:stefan at sun.ac.za>> wrote:
> >
> >     On Fri, Jan 05, 2007 at 09:38:49AM -0500, Neal Becker wrote:
> >     >     Several extensions to Python utilize the buffer protocol to
> >     share
> >     >     the location of a data-buffer that is really an N-dimensional
> >     >     array.  However, there is no standard way to exchange the
> >     >     additional N-dimensional array information so that the
> >     data-buffer
> >     >     is interpreted correctly.
> >     >
> >     > I am questioning if this is the best concept.  It says that the
> >     data-buffer
> >     > will carry the information about it's interpretation as an
> >     N-dimensional
> >     > array.
> >     >
> >     > I'm thinking that a buffer is just an interface to memory, and
> >     that the
> >     > interpretation as an array of n-dimensions, for example, is best
> >     left to
> >     > the application.  I might want to at one time view the data as
> >     > n-dimensional, but at another time as 1-dimensional, for example.
> >
> >     You can always choose to ignore that information if you don't need
> it.
> >     On the other hand, if you *do* need it, how would you otherwise
> >     interpret an N-dimensional array, given only a buffer?
> >
> >
> > I think Neal is suggesting some object that basically does nothing but
> > hold a pointer(s) to memory. This memory can be used in various ways,
> > one of which is to use it construct another type of object that
> > provides a view with indices and such, i.e., an array. That way the
> > memory isn't tied to arrays and could concievable be used in other
> > ways. The idea is analagous to the data/model/view paradigm. It is a
> > bit cleaner than just ignoring the array parts.
>
> Such an object would be useful.  I would submit that it is what the
> buffer object "should be"


Yeah. The problem is that we have a buffer API, not a buffer object.

Chuck
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://projects.scipy.org/pipermail/numpy-discussion/attachments/20070105/8738482d/attachment.html 


More information about the Numpy-discussion mailing list