Map of multigraded modules removes degrees

[mahrud]

Here is an example:

A = kk[x_0..x_1, DegreeRank => 2]
B = kk[a_0..a_1,y_0..y_1, DegreeRank => 4]
f = map(B, A, {B_0, B_1})
degrees A^{{1,2}}     -- {{-1, -2}}
degrees B^{{1,2,0,0}} -- {{-1, -2, 0, 0}}
degrees f A^{{1,2}}   -- {{ 0,  0, 0, 0}}, seems wrong
degrees sub(A^{{1,2}}, B) -- same

[DanGrayson]

If f were homogeneous, it would probably do something more sensible. Try giving f a degree map.

[mahrud]

Try giving f a degree map.

The usage of DegreeMap for multigraded rings is hard to use (and poorly documented), but regardless, I shouldn't need to. Here is another examples where M2 should just do the sensible thing:

A = kk[x_0..x_1, DegreeRank => 2]
R = A ** A
f = map(R, A)
degrees f A^{{1,2}} -- {{0, 0, 0, 0}}

S = A ** (kk[y])
f = map(S, A)
degrees f A^{{1,2}} -- {{0, 0, 0}}

[mikestillman]

I think that having it do something better by default is a good idea.

[mahrud]

Yes. Though both map(Ring, Ring, Matrix) and kernel RingMap are intimidatingly long for me. I think it might help if I broke them into smaller functions, perhaps using hooks to take care of special cases like maps involving Galois fields, then we can focus on the piece that guesses the correct degree map.

[mahrud]

What is a homogeneous ring map? The documentation doesn't say.