Re: [Axiom-mail] Reference on limits computations

[Martin Rubey]

An afterthought:

I believe it would be helpful to implement my favorite function hierarchy

(rational functions < algebraic functions < d-finite functions < differentially
algebraic functions < general function) 

and my favorite recurrence relation hierarchy

(linear with constant coefficients < linear with polynomial coefficients <
"pure" < general recurrence)

since for many of those classes specialized exact algorithms (for limits,
integration, summation,...) exist.  I'm buried in work, but if somebody would
seriously attack this, I'd be happy to cooperate.  Since I did already a bit of
work on these matters, I do believe I could contribute.

For me, the most difficult part is to get the "category"s designed right.  Note
that we want to cover not only the univariate case, but also the multivariate
case, the case of other Ore algebras, etc.

Martin