Re: [Cbor] An Information-based CBOR schema language?

[Carsten Bormann <[e-mail]>]

Hi David,

I cannot answer your question on the level of interest in doing work on another data description language, but I want to quickly address two statements in your email:

> On Aug 21, 2019, at 16:19, David Kemp <[e-mail]> wrote:
> 
> CDDL is a schema language for CBOR.   JADN (JSON Abstract Data Notation) is another.  JADN differs from CDDL in that it:
>   1) Specifies an Information Model rather than a Data Model (RFC 3444), and

The prose around JADN says so, but then goes ahead and does approximately what CDDL does (actually it appears to me that in some places it is even more concerned with serialization issues than CDDL is).

Since the boundary between information and data models is not strict (and, indeed, one may want to make several refinement steps on the way from a conceptual model to a data model that directly binds to a serialization scheme), this is not very surprising.

As one data point, I would like to point out that CDDL is very much in use as an information modeling language; this can be mostly done by stopping short of giving a specific mapping to JSON or CBOR constructs (or by using container/composition abstractions that only have default mappings, which enables the information model to be “executable”, without actually tying down the serialization mappings of those abstractions).

JADN does address one point CDDL doesn’t do very well yet: Having a single model that maps to both idiomatic JSON and concise CBOR.  I believe that the next version of CDDL will need to provide more functionality for this, as well.

>   2) Is itself structured data (JSON or CBOR) rather than text that requires parsing

Please see https://tools.ietf.org/html/draft-bormann-cbor-cddl-freezer-02#section-8 for why this may not be a distinction for much longer.  Actually, JADN also defines a human notation (JADN-IDL), and I applaud you for this decision.  (I don’t know why humans, outside of debugging, should ever need to see the JSON formats of either JADN or CDDL.)

In summary, I do believe that JADN is an interesting development that we as a WG should take note of.

Grüße, Carsten