Igor Schein on Tue, 10 Sep 2002 13:34:51 -0400

[Date Prev] [Date Next] [Thread Prev] [Thread Next] [Date Index] [Thread Index]

Re: yet another rnfkummer() posting

On Tue, Sep 10, 2002 at 07:30:45PM +0200, Karim BELABAS wrote:

> I believe this would make bnfinit() about 3 times slower for easy examples
> (say a few minutes running time), and it shouldn't be noticeable for tough
> examples. In short it should be quite OK for individual fields, but possibly
> annoying for large scale computations dealing with a lot of fields.

My preliminary vote is against the default slowdown.

> For direct calls to bnfinit, there's always the possibility of explicitly
> cheating by fixing C2, but for internal calls in high level fonction (e.g
> rnfkummer, bnrstark, etc), the user is stuck, unless we make C2 a 'default',
> but then we first have to make the library aware of defaults first [ which
> would not be a bad thing ]
> Still thinking...

Maybe make C, C2 global defaults?