On 3/21/07, Kyle Schalm <[EMAIL PROTECTED]> wrote:
> btw, i did have trouble applying the patch. it said,
> "abort: outstanding uncommitted changes".
> any idea what that means?
It means it's a text-only patch that was taken "out of context"
from my tree to yours. You can get around though -- do
On 3/21/07, Robert Bradshaw <[EMAIL PROTECTED]> wrote:
> I agree with Kyle that small cases should default to the naive
> algorithm, 2x2 at least, especially if working in the fraction field
> is expensive in comparison. Should I change this?
Yes, definitely.
I think the generic Hessenberg form
Here's a patch that resolves the underlying issue, namely that
elements of the fraction field over a generic polynomial ring
couldn't be cast back into the original ring despite having
denominator 0. This came up because the Hessenberg form calculation
involved passing to the field of fract
> On 3/20/07, Kyle Schalm <[EMAIL PROTECTED]> wrote:
>> there is trouble with the determinant method on a matrix over a funky ring
>> (yes, the same funky ring causing all my other problems). in its simplest
>> form:
>>
>> In [43]: W.=QQ['w']
>> In [44]: WZ.=W['z']
>> In [45]: matrix(WZ,2,2,[1,z
On 3/20/07, Kyle Schalm <[EMAIL PROTECTED]> wrote:
> there is trouble with the determinant method on a matrix over a funky ring
> (yes, the same funky ring causing all my other problems). in its simplest
> form:
>
> In [43]: W.=QQ['w']
> In [44]: WZ.=W['z']
> In [45]: matrix(WZ,2,2,[1,z,z,z^2]).det
