another two bugs in solver



Martin RUBEY <rubey@labri.fr> writes:

> Unfortunately, maxima starts to become unusable for me now.

On the other hand, SOLVE does work for the CRE form of the equations

(C1) eqs:rat([u-t*(u^(l+1)+u+1),1-t*((l+1)*u^l+1)]);
		     l				       l
(D1)/R/      [- t u u  + (- t + 1) u - t, (- l - 1) t u  - t + 1]
(C2) solve(eqs,[t,u]);
Is  l  an integer?

yes;
				   l
				 -----
				 l + 1
				l		  1
(D2) 		      [[t = --------------, u = ------]]
			       l		   1
			     -----		 -----
			     l + 1		 l + 1
			    l	   + l + 1      l


Actually, there's a bit of mathematical reasoning behind using RAT
here: Normally, when trying to manually solve this system of
equations, something like substituting a new variable for u^l
immediately comes to mind, but that's exactly what RAT is doing (that
is, u^l is a `kernel' of the expression), so...

Wolfgang