Actually, to get Hennessy's suggested behavior, it would be very simple to
have mapply barf if %void was given as an argument to a function.
Of course, that means that it would become impossible to test
is(limit(...)=inf), to return a list [limit(...), limit(...)], etc.
All in all, a terrible idea, but easy to implement!
-s
On Mon, Oct 5, 2009 at 10:04 AM, Richard Fateman <fateman at cs.berkeley.edu>wrote:
> what may not be clear to everyone in this discussion is that adding a
> new object with special properties like the ones suggested for "void"
> requires an examination of many places in Maxima.
> For example, the instructions for simplifying "void + x" would not be
> encoded in the object "void". These are encoded on the property list of
> the object "+".
> Thus to make "void", or for that matter, inf, und, ind, intervals, etc
> work, one has to modify +, *, sin, cos, exp, log, integrate, diff, etc
> etc etc.
>
> Now maybe this organization is wrong, but that's the way it is.
> (arguing about whether cos(void) should send a "void" message to "cos"
> or a "cos" method to "void" does not
> offer a solution to "void+infinity+interval(-1,1)")
>
> _______________________________________________
> Maxima mailing list
> Maxima at math.utexas.edu
> http://www.math.utexas.edu/mailman/listinfo/maxima
>