Ext into arbitrary target modules

[hoodmane]

Currently we compute Ext(M, k) from a (possibly derived) module to the trivial module. We should be able to compute Ext(M, N) for N an arbitrary module.

[dalcde]

At the moment there is one minor technical problem when dealing with the Hom of two chain complexes, which is that the minimum s-degree is now negative. This is not a major issue, but how do you intend to deal with that?

[hoodmane]

Does the code not currently allow negative minimum s degree? If there is some issue with negative min s, then first fix that issue. I haven't looked at the code for a while, probably you're upset that we use Vec rather than BiVec for the s direction? Another option is to translate the s degree to start at degree 0, but it would probably be better to use BiVec for s degree.

[dalcde]

It currently does not. Since you were halfway through that I thought you might have had some plans already. Rewriting everything so that s is an i32 instead of a u32 is probably a huge pain, but might be the correct thing to do. At the moment, the spectral sequence manager *does* support negative s degrees, so that's some work done already.

…

On Mon, Feb 10, 2020 at 08:30:25PM -0800, Hood Chatham wrote: Does the code not currently allow negative minimum s degree? If there is some issue with negative min s, then first fix that issue. I haven't looked at the code for a while, probably you're upset that we use Vec rather than BiVec for the s direction? Another option is to translate the s degree to start at degree 0, but it would probably be better to use BiVec for s degree. -- You are receiving this because you commented. Reply to this email directly or view it on GitHub: #27 (comment)