--- Richard Fateman wrote:
> If you want to save typing, try
>
> s(x):=ratsubst(1,u^2+v^2+w^2,x);
>
> then type
> s(%);
> every so often.
Okay, I guess this would work in this simple case, but
you could see where it would become almost as tedious
in complex situations.
>
> or you could define a rule and use
> tellsimp or tellsimpafter.
> Then it would be used automatically,
> but perhaps too much.
Is this what is going on inside the "trigsmp.mac"
file? I figured I could just copy whatever is going
on with sin^2 + cos^2 = 1, but after I looked at that
.mac file, I realized I'd have to read about 56
kajillion other such files in order to understand the
mechanism.
I figured it *had* to be simpler than that ...
>
> As for ::
> it is almost never used. It certainly
> doesn't do what you seem to think it does.
Well, if what it's doing now is rarely useful and not
very intuitive anyway, then I propose that this
operator be "hijacked" and reassigned to automatic
substitutions like the one I outlined in my original
post.
>
> Why not just set w: sqrt(1-u^2-v^2).
Again, this will work in my simple case where you can
easily solve for one variable explicitly in terms of
the others. You can see where this approach would not
work in general. I mean, are there not cases where
Maxima's "solve" facilities cannot convert an implicit
"function" into such an explicit form?
On the surface of it, it seems to me that getting
"solve" involved in a "subst" type operation
needlessly complicates the issue.
Thanks,
Dan
__________________________________
Yahoo! FareChase: Search multiple travel sites in one click.
http://farechase.yahoo.com