Clarify the atomic parts and data types of the model

[nichtich]

Clarify the atomic parts and data types of the model

[michael]

Yeah. Good points. And the answer is, yes. We're trying to pick a useful featureset (properties on document-level) an a sequence of nodes (that can be customized). I'm aware that the tricky part is finding the right balance between generalization and specialization. The substance document model is just an application of JSON, it defines the format and the protocol to manipulate it. I'm trying to reuse existing technology whenever it makes sense. Specifying document schemas using JSON-schema might make sense. However those parts don't have priority at the very moment, since the API needs to be solid at first.

We'll have a working demo soon, I'm quite happy with what we've got so far. Incremental document updates through operatoins give us a whole lot of flexibility.

Will keep you posted about progress.

[michael]

We're just working on a reference implementation here, supporting a bunch of common content types (text, section, image, list). So this will be a concrete instance of the Substance Document Model, that comes with a web-based editor to manipulate it.