Stavros Macrakis wrote:
> On Mon, Apr 14, 2008 at 11:10 AM, Raymond Toy (RT/EUS)
> <raymond.toy at ericsson.com> wrote:
>> I'd much rather see them all converted to member than leaving the code
>> in this half-finished state.
>
> If only this were the least consistent thing in our codebase! We have
> something like 9 different "standard" ways Maxima code performs a+b
> where a and b are in general representation:
> (simplify (list '(mplus) a b))
> (simplifya (list '(mplus) a b) nil)
> (simplus (list '(mplus) a b)) -- probably a bug: blocks pattern-matching
> (list '(mplus) a b) -- lets the simplification happen later
> (m+ a b)
> (add a b)
> (make-mplus a b)
> (lapsum a b)
> (poisco+ a b) -- probably justifyable so the representation of
> Poisson coeffs can be changed
>
> And no, I do *not* encourage anyone to go around standardizing these
> -- unless they're actually making major changes to the code to improve
> functionality anyway.
It's one thing to have inconsistencies that have been inherited, but
it's another when we intentionally introduce them. :-)
I think I have used both m+ and add in various places in the code,
because that's what they were doing. Is there a preference for one or
the other? We could standardize on one, if, in fact, they are the same.
Ray