DID applications should be able represent data in whatever manner they choose, whereas the representations of data in a DID document about the DID itself should be specified by the DID standard. To meet these dual requirements, these two classes of data must be syntactically differentiated.

We need to evolve the DID document syntax to syntactically differentiate between data about the DID versus data for DID applications that happens to be co-resident in the DID document. The former is the domain of the DID standard. The latter is not.

Note that this differentiation was initially discussed by @dlongley and @selfissued in #67.