Ben Thompson

For the most part, I think we should treat these as bugs unless there's something that I'm missing. But, I guess I have a general concern about the scale of...

Yes, I strongly approve of having a clearly documented API that is consistent between the various MatrixBase classes! Thanks for pointing this out, Marc. Part of the task here seems...

Marc and I talked a couple days ago about some indexing issues where the capabilities of DenseMatrix greatly exceeded that of other matrix types because the underlying numpy arrays have...

I think we should distinguish between a fuller API and an inconsistent API. Fuller is fine but inconsistent is not. Does that seem reasonable? And like the title of this...

See #107 and #108

And remove Python 3.6 support.

Relatedly: https://github.com/flatironinstitute/sparse_dot/issues/9

Yes, you're correct that `_split_col_subsets` currently assumes the input columns are ordered monotonically. One thing we could do is sort the columns first, then call `_split_col_subsets` and then "unsort" the...

I actually ran into this same issue from the other "end" recently when I fixed some bugs with subsets that were not monotonically specified. So this same bug is a...

@MarcAntoineSchmidtQC did this get fixed or did it get dropped when you closed your PR? Is it worth reviving just this fix?