Thank you very much, the charpoly doesn't abort.
There's a minor catch, though.
The result is virtually unreadable as a whole,
can't be parsed(?) or indexed(?), and there
seems to be no info on degeneration, anyway?
-Yury
On 07/28/2013 07:25 PM, Leo Butler wrote:
> From: Yury Tarasievich <yury.tarasievich at gmail.com>
>
> Hi guys,
>
> On Maxima 5.30/linux/amd64, I'm getting this
> with fairly simple block-diagonal symbolic
> matrix of 10x10 and/or 11x11 dimension:
>
> E:eigenvalues(opH)
> part: fell off the end.
>
> But with matrix of dimension 9x9 and lower the
> routine completes okay.
>
> What could/should I try? Should I post the
> matrix in the form it's 'print'ed, too?
>
> The alt_eigen solution recommended earlier in
> this list can't pass its own tests on 5.30.
>
> Is there anything else?
>
> Yury, have you tried
>
> p : charpoly(opH,x);
>
> ?
>
> You likely have a factor in p that is of degree >4.
>
> Leo
>