#5 starting here (I'll call this "Ian's summary") proposes an abstract syntax (used synonymously with 'data model') for Ocapn. Together with subsequent discussion in that issue and perhaps elsewhere, we may have consensus to a large extent on what the abstract syntax should be in the eventual Ocapn specification.

I would like for the current consensus on the abstract syntax to be captured in some living document we can use for reference in implementation, testing, and specification work. There will be fine points on which we do not yet have consensus, which can be called out in the document. It's fine if the document borrows significantly from Ian's summary and the rest of issue 5; it does not have to be 'publication quality'. Ideally at our November meeting we can call for consensus on this document.

Concrete syntax is out of scope here. See #93