On Oct 30, 1:14 pm, "Mike Hansen" <[EMAIL PROTECTED]> wrote:
> When I said recursively defined streams, I meant something a little
> more general.  For example, if f is a power series and g is exp(f),
> then g satisfies g = \int g*f'.  This translates to the following code
> for the exponential of power series:
>
> g.define( (f.derivative()*g).integral(base_ring(1)) )

If that works, the better.

>
> The benefit of this is that it reduces the complexity of
> exponentiation to that of multiplication for which there are fast
> algorithms.  Generally, you want to reduce everything (composition,
> division, logs, etc) to multiplication in order to get asymptotically
> fast algorithms.

Sure, but I would say functionality before performance.

> > What about lazy Laurent series?
>
> Are you asking how one would implement them?  What sort of features
> would one require for lazy Laurent series?

I am asking: why not more generally implement them instead of purely
powerseries, I think they are only a slight generalization of
powerseries.

> > What about moving from combinat to rings?
>
> The reason why it was in combinat/ is because I had written it for my
> particular use case.  There are some things that need to be done for
> it to be usable in a wider setting.  For example, certain things only
> make sense when you have a power series over a field.

I was also thinking about similar issues. Also some operations require
and guaranty the first coefficient to be 0.
I wonder whether to introduce a subtype for those powerseries because
it is not always decidable whether something is 0 in the base_ring.
Formal Laurent series would also form a field.
For example the formal Laurent series are a field.
--~--~---------~--~----~------------~-------~--~----~
To post to this group, send email to [email protected]
To unsubscribe from this group, send email to [EMAIL PROTECTED]
For more options, visit this group at http://groups.google.com/group/sage-devel
URLs: http://www.sagemath.org
-~----------~----~----~----~------~----~------~--~---

Reply via email to