[GiNaC-list] Silent cancellation of singularities, and others

Alexei Sheplyakov varg at metalica.kh.ua
Wed Aug 12 16:28:01 CEST 2009


Hi,

On Wed, Aug 12, 2009 at 02:45:07PM +0200, Burcin Erocal wrote:

> I thought that caching might help with printing large expressions
> repeatedly. It could well be that it is pointless.

I doubt someone will want to look at large expressions except for debugging.
And for debugging one needs a printing function which is fast, simple, and
free of side effects (such as sorting the sequence).

> I was hoping that the capability to work with infinity would help
> series expansions, where 1/ (1/0) could be evaluated as 1/infinity -> 0.

I *really* dislike when singularities get silently canceled. Also, is it
+0 or -0? 0 + i 0 or 0 - i 0? (these details are important more often
than not).

> Especially, automatic simplification of powers and multiples of exp
> 
> http://pynac.sagemath.org/hg/rev/24e8ecd16228 exp(a)*exp(b) -> exp(a+b)

The code looks correct. But I dislike random features like this.

> http://pynac.sagemath.org/hg/rev/5c4862f90e17 (e^x)^y -> e^(x*y)

I think it can be trivially done with pattern matching.

Best regards,
	Alexei



More information about the GiNaC-list mailing list