We already have a number of issues related to this, but we never set out to do a document to clarify what we want/need, hence this issue:
Studying tskit code I see that when exporting to Nexus, Newick, or FASTA files, it uses n0, n1, ... type naming for node IDs. Not sure if relevant for us.
Useful docs on requirements to get a valid tree sequence: https://tskit.dev/tskit/docs/stable/data-model.html#sec-valid-tree-sequence-requirements
@LynxJinyangii has already done quite a bit of exploration and prototyping at https://github.com/LynxJinyangii/AlphaSimR/blob/master/dev/notesRealBreakpoints.md
We effectively want something like what Ben Haller did with his document https://github.com/MesserLab/SLiM/blob/master/treerec/implementation.md, but that document merges the initial white paper (=plan) and later implementation, so we will end up with such a doc at the end, but we need a clear plan first and then implementation.