Categories of software: inference vs simulation

[hyanwong]

We used to be able to categorise the list of software projects into "simulation", "inference", "base", etc. I think this is a useful thing: it is handy to see e.g. the diversity of software that can be used to simulate tree sequences. This is rather orthogonal to the current classification into tskit-dev projects and Ecosystem projects. Are we thinking of being able to bring the classification back somehow? It's useful to be able to link e.g. to "software that infers tree sequences", and "software that simulates tree sequences".

One possibility is to add a separate list at the bottom of the "software" page. That way we could include software in the list which are neither tskit.dev nor ecosystem projects, such as Relate and KwARG, but which have (or will have) converters that mean that their output can be turned into tree sequence format.