CoCoA » CoCoALib

Here is the background for my request.

I want to modify MinPolyQuot so that it accepts "relatively zero-dim ideals": e.g. an ideal whose RGB (which has already been computed) involves only x and y even though the ring is QQ[x,y,z]. What I am thinking of implementing is: find out which indets do not appear in the RGB, then create a new RGB being the old one with each "missing" indet added as a new RGB element.

I suppose I could try creating a new poly ring (with fewer indets); map the computation into this smaller poly ring, and map the answer back. To do this cleanly, I'd need to create the new poly ring with an ordering which is compatible with the original one (but restricted just to those indets which actually appear), and then create an ideal in this new ring; but again I want to be able to say that I know that these gens really do form a (reduced) GBasis.

I was aiming to use the first approach because it seemed easier to implement. The second approach is perhaps mathematically cleaner?

There are already fns IdealOfGBasis and SetGBasisAsGens. Where is the doc for the 2nd fn?

Despite some initial doubts, I now think that this issue is genuinely different from issue #1255, since here the idea is to trust the caller blindly, whereas #1255 would imply having to do some (possibly quite a lot) computation.