applying function identities (proposal)



----- Original Message ----- 
From: "Robert Dodier" <robert.dodier at gmail.com>
To: "Barton Willis" <willisb at unk.edu>
Cc: <maxima at math.utexas.edu>
Sent: Tuesday, May 02, 2006 7:39 AM
Subject: Re: [Maxima] applying function identities (proposal)


> Hi Barton,
>
>> Two defrules can display the same, yet be different:
>
> Yes, this is a shortcoming of the current system.
> However it's easy to remedy -- just carry the matchdeclare
> info along with the rule definition and display the rule as
>
>  r1 : f(x) -> x f(x - 1) assuming mapatom(x)
>

This may not solve the problem!  say you apply rule r1, and
then change the definition of mapatom, and apply rule r1 again.
The rule will have 2 different meanings.

The binding of the name to the function happens at the last possible moment,
I think.

I'd prefer

   matchdeclare(x_atom, mapatom),

as I said in the previous message (that crossed in the email..)

RJF