> Really, what a robust commentary system needs is to map many comments to many units of text This is actually built into this specification. From the Web Anno

> Really, what a robust commentary system needs is to map many comments to many units of text This is actually built into this specification. From the Web Annotation Data Model [0]: - Annotations have 0 or more Bodies. - Annotations have 1 or more Targets. So one "Annotation" object can have multiple bodies (descriptions) attached to multiple targets. > 3. Relationships between comments This sounds more like an implementation detail of a client than part of the protocol or data model put forth by the W3C group. However, I believe this can kind of be done server-side with the Web Annotation Protocol [1]'s idea of Annotation Containers. Your server can map a single annotation to multiple containers. So perhaps you have an endpoint like `http://example.com/a/` http://example.com/a/` and you want to arrange a hierarchy of comments. You could provide a filtered set of the annotations at `http://example.com/a/romeo/consonance/` http://example.com/a/romeo/consonance/`, and similar endpoints. So basically what I'm saying is it seems like the protocol here isn't going to get in your way, it's just incentive to use this particular model for storing and transferring your data. [0]: https://www.w3.org/TR/annotation-model/#web-annotation-principles https://www.w3.org/TR/annotation-model/#web-annotation-princ... [1]: https://www.w3.org/TR/annotation-protocol/#container-retrieval https://www.w3.org/TR/annotation-protocol/#container-retriev...