[PEP draft 2] Adding new math operators

[Robin Becker]

In article <slrn8p0qmg.14e.hzhu at rocket.knowledgetrack.com>, Huaiyu Zhu
<hzhu at localhost.localdomain> writes
...
I strongly disagree that we have a clear winner. That is an assertion
with no evidence except your statement. I personally think the use of
concatenated operator symbols may lead to many difficulties, but I
recognize that it is an opinion only.

I think that python will suffer from the addition of another bunch of
parallel operators and the meta method names to go with them.

Before deciding on this approach we should at least determine how often
these operations are actually mixed. I very rarely use the element wise
operations with matrices, but others may. We should at least gather some
information before making a complexity change of this magnitude. Among
other things whether ~ is a good start op candidate when it is already
in use is doubtful.

Another thing that should be borne in mind is that we'll be moving to
unicode so that other operator notations may in fact be more feasible in
future eg the symbol character set may have an existence in unicode.
That would make the symbology easier, but not how to associate such
symbols with python code.
-- 
Robin Becker