Oscar Benjamin

I think that the next thing that is needed is a function that can take the output of `factor_system` etc and turn it into a system of non-redundant irreducible Groebner...

Actually I think that `factor_system` and `factor_system_cond` should be changed. I realise that this case is not handled well by `factor_system`: ```python In [34]: factor_system([a*x, y, a]) Out[34]: ([[y, a]],...

> So I think what should happen maybe is that if there is a symbolic "constant coefficient" like `a` then it should just be multiplied back in to each factor....

> > So I think what should happen maybe is that if there is a symbolic "constant coefficient" like `a` then it should just be multiplied back in to each...

> > > So I think what should happen maybe is that if there is a symbolic "constant coefficient" like `a` then it should just be multiplied back in to...

I think that it would simplify everything to treat the constant factor as just being one of the factors rather than separating it out. I tried this out with: ```python...

Yes, let's continue the work here.

> Shall we rather utilize `factor_system_cond`'s output here ? It should be possible to work with the output of either. If `factor_system` is used then there won't be any systems...

I think it might be better to have two subclasses of PolyRing and then have `PolyRing.__new__` return the one or the other.

The PR gh-28110 is doing too many things at once. It is going to be impossible to review by changing everything in one go. I think it is better to...