Assuming I get it included, is there any features in particular you'd want to see in there? Note that if I do have it produce visualisations, they'll be static images as part of an analysis report rather than being interactive.
I was just recently wondering about something like this. I'd like the ability to show individual module dependencies, and then to collapse modules in one package to just the package, so I could "zoom out" and see how the packages relate to each other. By "package" here I mean the "A" in "A.B, A.C, etc." It seems like it would be fairly simple to use Language.Haskell.Parse to turn a set of modules into a graph, and then something to massage that and give it to dot. It actually brings up a question about module discipline in general: At one point I wanted to have each package re-export its public modules and then sibling packages can only export the "package" module as some libraries do, but it got to be too much re-exporting bookkeeppiing and I didn't so much like symbols having multiple "homes", and most of all it was increasingly awkward to split packages into strictly higher level and strictly lower level relationships since this discipline means coarser granularity which means more trouble avoiding circular imports. So now I just import modules directly from their packages. I still have a general structure of higher, lower, and sibling packages though, so some kind of visualization would be handy to keep deps under control, look for packages and modules that could be split up for a clearer hierarchy, find good places to cut circular imports, etc. I've also thought I could move private modules in A to A.Private, so you can see that if you're importing X.Private.Y from a module that doesn't start with X. then something fishy is going on. How do other people approach this issue?