Am Sonntag, den 19.09.2010, 12:53 -0400 schrieb Stavros Macrakis:
> Unfortunately, Maxima's core inequality system is not very powerful,
> and cannot conclude from H>0, d>0, H>d that H^2 > (H-d)^2.
>
>
> Barton Willis's fourier_elim package is much more powerful, but
> unfortunately is not currently integrated with the 'assume' database.
> It would be good if we could integrate fourier_elim into the core
> inequality system -- the main drawback is that sometimes fourier_elim
> can be very computationally expensive.
In addition to the weakness of the assume database we have a bug ID:
2872553 "Assume database inconsistent after integration". I did not look
into the problem. Perhaps it is a related problem.
Dieter Kaiser