Hi again!
Thanks for the help.
I had already reimplemented "triangularize" in Maxima
and my implementation seems much faster at least for
integer coefficients than the built-in version.
Maybe it can be useful to someone.
I guess the speed-up is due to the fact that I used
arrays instead of lists.
The algorithms uses division and no smart pivoting.
So there is lots of room for improvement.
It will be part of a package for real algebraic geometry
which I will submit hopefully this summer or a bit later.
Regards
Fabrizio