On Thu, Jan 5, 2012 at 11:18 AM, Douglas Burke <[email protected]> wrote: > On 1/5/12 10:16 AM, Diab Jerius wrote: >> >> On Wed, 2012-01-04 at 21:36 -0500, chm wrote: >>> >>> I propose we use "shape" as the standard term to >>> describe the set of dimension extents for a piddle. >> >> >> Excellent idea. >> >> >> _______________________________________________ >> Perldl mailing list >> [email protected] >> http://mailman.jach.hawaii.edu/mailman/listinfo/perldl > > > It may be well to use terms that are the same as those of other > computational systems, to make it easier to swap between them/transcribe > code/argue over which is better/... > > However, NumPy uses the term rank to give the number of axes (or dimensions) > to an array, and shape as the specific size/number of elements/.. of an > array, as described at > > http://www.scipy.org/Tentative_NumPy_Tutorial#head-6a1bc005bd80e1b19f812e1e64e0d25d50f99fe2 > > which use appears to be dis-favored by the discussion so far ;-)
Yes, their use of shape is consistent with this proposal. I prefer use rank in the linear algebra sense for which this use is only the same if the matrix has a null nullspace. > Not sure what other systems such as octave use, > Doug _______________________________________________ Perldl mailing list [email protected] http://mailman.jach.hawaii.edu/mailman/listinfo/perldl
