Re: [Maxima-discuss] order of assumptions matters

[Robert D. <rob...@gm...>]

On Tue, Feb 23, 2021 at 5:15 PM Barton Willis via Maxima-discuss
<max...@li...> wrote:

> I've been a maxima user for a few decades, but only now did I discover this.  So, no fixing this isn't a high priority.  But replacing the algorithms behind assume, sign, and such would be a lovely SMOP activity.

The assume database is not too complicated in its basic form. It
appears to be an ordered heap. For assertions about one variable it
seems straightforward enough; I think where the ordered heap is not
really up to the job is where there are two or more variables
involved. I don't know what kind of structure would be appropriate for
that; any ideas on that point would be most welcome.

The programming around the assume database is a nightmare, but it is
pretty well contained, so a drop-in replacement seems plausible. We
could probably fix dozens of bugs at one fell swoop.

I would be interested to hear about what approach is employed by other
systems such as Reduce or Sympy.

best,

Robert