James Fairbanks

I think that as long as we are iterating over the paths and not realizing them all in memory, it is a fine addition. I draw that distinction because users...

I think including it with that complexity warning in the docstring is the right move. Since the algorithm can change without breaking the interface (when you iterate the iterator you...

I think I understand Seth's point to be distinguishing the weights object from the weighted adjacency matrix. If you do collect(weights(mg)) .* adjacency_matrix(mg) you get the weighted adjacency matrix. A...

I think code that uses these fancy matrix multiply algorithms will require a special skill set for the maintainers and don't think that LG can provide that level of expert...

What about the case where sigma is negative because of integer arithmetic wrap around? Do you still want 0 closeness?

An extra branch that won't get hit except in exceptional circumstances seems worth it to me. Well, can we prove that this can't overflow because it is a sum of...

I think the right answer here is to call `HTTP.Strings.escapehtml` on every vertex and edge property before we write it to the IO buffer.

This is a good thing to figure out as having Pluto work with Catlab would be really awesome. What would we need to do to make it Pluto-friendly? We have...

Yeah, see my latest PRs #563 and #566. I am adding one file of docs and expect that everything else is going to be fine. My current workflow is to...

We just got bit by this on Decapodes. Its really a problem when we have theory inheritance. For example making a syntax for the Exterior Calculus requires importing a lot...