[MATRIX-SIG] A proposal (LONG) - "reverse of take" revisited

Phil Austin phil@geog.ubc.ca
Tue, 1 Jul 1997 18:54:04 -0700


    Aaron> I'm reeling about not following this discussion, but I
    Aaron> suspect the best approach is for Jim H. to "just do it".  I
    Aaron> personally begin to think that excess overloading of
    Aaron> indexing is not advisable, possibly confusing.  I vote for
    Aaron> "take()" as is with dual "put()" or whatever.  Of course,
    Aaron> my experience with numeric matrix tools such as these is
    Aaron> infinitesimal, so my opinion may be of limited merit.  If
    Aaron> we can get by without asking Guido to change the core, so
    Aaron> much the better. -- Aaron Watters

Here's the perspective of someone who intends to teach NumPy to 3rd
year geography majors without much exposure to computers: I would
<much> rather have the S syntax describe by PSC and APM than the
current numeric functions.  Like Python, it takes a while to
penetrate, but once it does, even naive computer users typically find
it intuitive.  Additional evidence is available in the S-news
archives, where hundreds of real neophytes have asked questions of the
old timers. A couple of array examples are typically enough to get the
beginners started (and there are some good public domain S tutorials
that could be pirated for this).  Once the NumPy prototype code is
available, the indexing should sell itself to members of this list and
to the broader community.

Phil

_______________
MATRIX-SIG  - SIG on Matrix Math for Python

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