Igor Schein on Mon, 25 Sep 2000 17:15:40 -0400


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

Re: power series bug


On Mon, Sep 25, 2000 at 03:36:57PM -0400, Ilya Zakharevich wrote:
> On Mon, Sep 25, 2000 at 02:18:30PM -0400, Igor Schein wrote:
> > ? (acos(y)+z)^1.;
> >   ***   Warning: normalizing a polynomial with 0 leading term.
> >   ***   Warning: normalizing a series with 0 leading term.
> >   ***   division by zero in gdiv, gdivgs or ginv
> > 
> > If I pre-introduce variable z to the stack, it works fine:
> > 
> > ? z^2;(acos(y)+z)^1.;
> > ? 
> 
> This is easy, even I can answer it: to take a fractional power of a
> series, you need to have the leading term of degree 0 - and with a
> coefficient you can take a power of.  The leading
> term of the second one is acos(y)*z^0 (it is a series in z), the
> leading term of the first one is z*y^0 (it is a series in y).
> 
> Applying the same argument, it makes sense to take a power of acos(y),
> but not of z.

Sorry for temporary brain eclipse.

Thanks

Igor