The current iteration of the HSDS Workgroup has been operating under the assumptions we are developing HSDS 2.1. However, there are breaking changes, whereby if we decide to stick with semantic versioning standards, we should be working on 3.0.
There is a desire, I think, to not scare the broader community with a new major version release, so I concede that perhaps we should ignore semantic rules and continue with 2.1 for optical reasons, but that should be determined in writing so that the decision can be referenced.