Vincent Traag

> The docs state a complexity of O(V^5), without reference to a specific algorithm. Just noting here that `igraph` does a simple loop over all pairs of vertices, checking `igraph_st_vertex_connectivity`,...

> `igraph_i_vertex_connectivity_undirected()` is quite wasteful in the sense that it first converts the graph to a directed one, and then runs the s-t connectivity on all i-j pairs, even though...

This seems to be the most recent reference on the topic that I can find: https://dx.doi.org/10.1145/3406325.3451088. This reports a runtime that is similar to the runtime of the maxflow algorithm...

> * This is supposed to be a very efficient max flow implementation: https://github.com/carlhub/hipr See this note in the `igraph` [documentation](https://igraph.org/c/doc/igraph-Flows.html#igraph_maxflow): > Time complexity: O(|V|^3). In practice it is much...

Shortest path functions in principle work OK with `NaN` values. That is, they do not crash. However, they are not consistent in their output and handling of `NAN`. I use...

OK, I now identified some problems. `igraph_vector_min` (and presumably some other functions) will return `nan`. Hence, checks on positivity for example do not work correctly when `nan`'s are involved. There...

No, I think that making changes to `igraph_vector_min` would not be the right direction. There are some other ways of checking the negative entries though, perhaps doing it manually. I...

Note to self: some other vector functions should also be checked still. Most notably `igraph_vector_(which)_minmax`, but perhaps also some others that should work correctly for NaN input still.

Are the subsets non-overlapping? F that is the case, you could map each node of a subset to an integer, indicating which subset it is part of. You could then...

I indeed considered that too trivial to implement in the C core. But if you feel it is better implemented in the C core, that would be fine as well....