Simon Schölly

I haven't had time to refresh my memory on min-cut algorithms, but I was wondering, was the old algorithm just incorrect for weighted graphs or are there also unweighted graphs...

Thanks for your contribution, I think we should indeed change how this function works for undirected graphs. We probably also need to adapt the tests (that are currently failing) and...

Thanks for your PR :) I added a few comments.

Better support for metadata should definitely be a long term goal. - About creating a subtype of `AbstractGraph`: - I am already doing some experiments with some abstract interface in...

The problem this benchmark is, that you only benchmark if for `p=0.5`, which is rather unrealistic for larger graphs. The implementation we currently have in Graph.jl, can certainly speed up,...

I mean, if we have some benchmarks, that show that this implementation is also faster for very small and very large `p` such as `p = k/n` and `p =...

It is indeed a bit unfortunate that this method returns sets of edges (which may not even have the same edge type as the input graph) and not sets of...

That thing with the `list` might have been inspired by networkx. I would agree, that one should be able to generate the neighbors implicitly. I would even say, that the...

Is it a bug though? Not rather a new feature? I fully agree with your first point, there is also an older issue for that:: https://github.com/JuliaGraphs/Graphs.jl/issues/98 About the second part,...

@etiennedeg @gjh6 I think this is something useful to have. Given that this PR has been lying dormant for quite a while, there might be some things that you would...