Re: Failure to simplify trig functions of simple rational multiples of %pi

Dan Hatton <[email protected]> writes:

> On Wed, 26 Aug 2009, Martin Rubey wrote:
>
>> The short answer is, unfortunately, "no".  Axiom (FriCAS, OpenAxiom)
>> currently does not support "assumptions" like "assume that l is an
>> integer".
>
>> There is (currently) no such type as "symbolic integer".
>
>> There has been a lot of discussion on how to implement what you want,
>> but as far as I know, nobody is working on it.  Would you be interested?
>
> Hmm... I don't (yet) know any lisp.  

(almost) no mathematical code in axiom (etc) is written in lisp, it's
all SPAD (a predecessor of aldor), which is almost the same language you
use to communicate with axiom as a user.  To see some programming
examples, look into the src/algebra directory of your distribution.

> For my own needs, I could probably hack something up with pattern
> recognition rules analogous to those in section 0.1.8 of Jenks and
> Sutor - but I'm not sure if I could give it sufficient generality to
> be any use to anyone else.

Yes, I guess so.  Just in case, *if* you want to get into assumptions
seriously, talk to Waldek and start a good design (probably not
easy...).

All the best,

Martin