[PYTHON MATRIX-SIG] alpha2 and default axes

Jim Hugunin hugunin@mit.edu
Tue, 13 Aug 1996 16:18:39 -0400

Thanks to all for your bug reports.  There were actually less than I
expected, I hope this is a comment on the quality of the code, and not
the quantity of the testing.

I want to get a new release out this week with all these little bug
fixes, but I'd prefer to address one major issue before then.

What to do about default axes?

I'd really like some more feedback from you people out there as to what
the default axes for the various NumPy functions should be.  The basic
functions I see here are:

ufunc.reduce, ufunc.accumulate, sort, argsort, argmax, ...
take, choose, repeat, compress

My current plan is to make the default argument for everything -1,
except use 0 for concatenate.  So far the only other proposal is from
Konrad and it suggests using 0 as the default argument for everything
except fft.  I disagree with this, but could easily see keeping 0 as the
default argument for take, choose, repeat, and compress.

More opinions would be helpful - Jim

MATRIX-SIG  - SIG on Matrix Math for Python

send messages to: matrix-sig@python.org
administrivia to: matrix-sig-request@python.org