Re: [Numpy-discussion] The NumPy Fortran-ordering quiz

[Charles R H. <cha...@gm...>]

On 10/18/06, Travis Oliphant <oli...@ee...> wrote:
>
> Charles R Harris wrote:
>
> >
> >
> > On 10/18/06, *Travis Oliphant* <oli...@ee...
> > <mailto:oli...@ee...>> wrote:
> >
> >     Charles R Harris wrote:
> >
> >     > Well, I knew that for numeric, but it was a good deal less
> >     obvious in
> >     > combo with the order keyword.  For instance, contiguous could
> change
> >     > its meaning to match up with FORTRAN, so that FORTRAN=True and
> >     > CONTIGUOUS=True meant Fortran contiguous, which was sort of what
> >     I was
> >     > thinking. Explicit never hurts.
> >
> >
> >     Ahh, so you were confused by looking at flags for 1-d arrays (where
> >
> >
> > More confused by what seemed to make sense to me. Tim was the one who
> > actually ran the experiment to see what was going on, and he wasn't
> > sure what FORTRAN meant either. Now if I had accessed the full set of
> > offsets, strides, and counts, it would all have become clear. Numpy
> > tries to hide the nasty details, which is good until you really have
> > to know what it happening underneath.
>
> I've done as you requested.  I've added F_CONTIGUOUS as an alias to
> FORTRAN and C_CONTIGUOUS as an alias to CONTIGUOUS.  These are the names
> that show up when you print the flags and you can use them whenever you
> used the other names.  The other names are still everywhere in the code,
> but perhaps the existence of these aliases will help people understand
> what is meant better.


Thanks, Travis. I really appreciate your willingness to make such
modifications.

Chuck