[Numpy-discussion] Latest Array-Interface PEP
oliphant at ee.byu.edu
Fri Jan 5 10:55:00 EST 2007
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
>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.
The simple data-buffer interpretation is still there. You can still
use simple "chunk-of-memory" only interpretation of the buffer. All we
are doing is adding a way for applications to ask if the object can be
interpreted as a strided N-dimensional array of a particular data-format.
So, this proposal does nothing to jeopardize the
buffer-as-an-interface-to-memory only model. I'm only using a table of
funciton pointers which is already there (tp_as_buffer) rather than
request an additional table of function pointers on the type object
(tp_as_array). I see the array view idea as fitting very nicely with
the buffer protocol.
More information about the NumPy-Discussion