[PYTHON MATRIX-SIG] Some thoughts on default axes.

[Jim Hugunin]

Konrad Hinsen wrote:
> 
> Jim Hugunin wrote:
> >
> > I really don't think that a complete set of both function should be
> > provided.  The great value in this proposal is it lets me define the
> > "proper" default axis for each function without fear of people not being
> > able to remember which one that is.
> 
> That supposes that the naming convention is easier to remember than
> the default axis, which is not obvious to me. I can imagine people
> wondering whether the correct function is fft or fft_.

I'd argue here that if you can remember that fft operates on the -1
axis, you can remember that all functions that operate by default on the
-1 axis have an underscore after them (which by the way is not a naming
convention I find reasonable).

What I like about this is that if you forget that fft operates on
axis=-1 by default you get an exception (for example the name fft is
undefined).  Otherwise, you instead get a perfectly legal array back
who's values are completely different from what you expected.  I find
these bugs extremely hard to track down (from personal experience).

Your concern about people wondering whether it's fft or fft_ is exactly
my concern about people wondering if its axis=-1 or axis=0.

> > If the names of these functions somehow made their categories obvious,
> > then I'd be happy to set the default axis for the nonstructural
> > operations to be -1.
> 
> I can't think of a naming scheme that would work like that.

I can't either, but I do think this would be a fine solution if somebody
can come up with the right names.

-Jim

=================
MATRIX-SIG  - SIG on Matrix Math for Python

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