On Sat, 2003-04-26 at 10:53, Richard Fateman wrote:
> are you talking about the common lisp package system
> or a new user-accessible package system for maxima?
Good question. I am talking about using the common lisp package system
internally in Maxima. I think that is also what Ray was talking about. I
am not sure about Martin.
> The latter would probably require mimicking the common
> lisp package systems (or perhaps a better version of
> such a thing) with naming conventions, import/export,
> hiding, etc. There is a package system in Mathematica
> which has generally been painful for the novice, and
> perhaps useful to the expert. I think this is similar
> to the common lisp situation.
I think that a Maxima-level package system is worth considering, but
should be considered a separate issue from the internal-lisp level
package system. Some entangling between the two is probably inevitable,
but I think the lisp-level packages can be considered before embarking
on Maxima-level package scheme.
> I also wonder how one deals with the properties of an old
> symbol that the user redefines. Does that go in a new package?
> How?
> Tellsimp rules for built-in functions fit in this category, I
> think.
I think the redefined rule would go in the maxima-user package. The
original would remain in the maxima package. Kill would remove the
maxima-user version and we would revert to the maxima version. I'm
saying this off the top of my head, so I may be missing something.
--Jim