CY,
Thanks for your comments.
> > Or perhaps some of you are "plain ASCII text" purists. Sigh.
> The few, the proud.
Throwing flamebait: my bad, please ignore. (Actually a worthwhile
discussion, but *not* on the Maxima list.)
> Actually, this explanation of ev might make a great deal
> of sense in the Maximabook
You may have noticed that many of my responses to users' questions are
much longer than the original question really demanded. That's because
I think of it as material for either user documentation or a design
issues notebook. So yes.
> If ev survives longer than it should
It already has, and I doubt it will ever get taken out. It will
certainly need to survive for backwards compatibility, even if
superceded for the command line "," syntax, and deprecated in general.
So yes, this description should go into the Maximabook.
> I'd never be able to describe it so well.
I'm sure you would, if you'd wastedXXX invested as many hours in the
writeup....
> We should probably do something about this. I remember some
> previous discussion about ev being somewhat messed up, and if
> we are going to fix it it would be very handy for
> documentation purposes if this were fixed sooner rather than later.
Cleaning up EV itself is futile. It is inconsistent and complicated by
design. "In general, it does what you want, unless you want
consistency." (Larry Wall) And some people like it; after all, there
are even people who like perl. (Flames in private email only, please.)
> Surely we don't want to require ev(''%,numer) though?
That is exactly the kind of issue that needs to be thought about.
And once you start thinking about the quoting/evaluation issue, you
realize that there may be other aspects of Maxima's evaluation and
simplification semantics which may benefit from rethinking.
-s