There have been a few fairly big loose ends to wrap up for Haystack 4. This is a broad suite of proposals to address those issues and modify a couple design patterns used in the Haystack ontology. These changes do include breaking changes in how defs are declared. But this proposal does not change how instance models are defined.

I've broken the overall proposal up into several sub-proposals:

1. 935: simplify how choices work for the common case to so that they work like all other tags
2. 936: redesign how choices work for the complicated specialization use case via a new design pattern using the of tag
3. 937: redesign ref of (domain/referent of the relationship) to use filters and eliminate the fluid-xxx def explosion
4. 938: finalize the question mark inference operator in filter syntax
5. 939: remove the equipFunction taxonomy until we learn more

Overall I think these are a great set of changes that simplify and complete Haystack 4.

I have reviewed your proposed changes with the team, and we like the direction this is taking, for sure.

In particular, the simplification of choices in 935 makes the creation and management of marker choices easier, eliminating unnecessary complexity.

And the inference operator for enhanced filtering (938) will allow for separating out equipment or not, without having to know which case you are dealing with (an ahu-equip with fan points, .vs. an ahu-equip and a fan-equip separately, joined by equipRef), as just one example.

Thanks Brian!

Eric