I’ve been toying around with using Alloy for understanding complex interactions between programming language features (it’s been fun so far!). Typical programming language constructs have lots of “contains” relationships (e.g. a type is defined inside a module).
Uniformly representing all of these using arrows gets messy in the visualizer, and it takes a bunch of hovering + projections to better understand what is going on. Here is an example:
The details aren’t super important, but the point is that some of the arrows are about contains relationships. For example, if I could have a big box with the label Crate0, that box would contain the boxes for T0, impl[Σ[α..ω]]1. Similarly for Crate1 and so on.
At the same time, just selectively hiding the arrows is not a solution (I’ve already hidden the non-essential ones above), because it is hard to understand the relationships otherwise.
GraphViz (for example) supports grouping nodes together using the subgroup construct (unfortunately, I don’t think it is flexible enough to support intersection between groups, but that’s OK for my use case).
Is it possible to do something similar in Alloy natively?
If not, any suggestions for how to develop a custom visualization? I took at look at Sterling which looks like it consumes XML generated by Alloy. (via @shriram’s post: Visualization for Alloy — what do you want?). I suppose it should not be terribly hard to set up a file watcher + reloading script where I consume the XML and have it regenerate a custom diagram in a separate window; this would avoid having to dive into internals of the Alloy visualizer.
