[apps-discuss] JSON Content Rules and the need for schema languages for JSON

[Andrew Newton <[e-mail]>]

All,

Since the topics of JSON Schema and schema languages in general are
being debated here, I thought I'd go ahead and publish a draft idea I
had for formalizing expected JSON structures. The draft is here:

http://www.ietf.org/id/draft-newton-json-content-rules-00.txt

While the notion that this is better (or worse) than JSON Schema is
subjective, I do think it turns out to be more concise.

Also, I think it worth mentioning what the ALTO working group has
done. They have a protocol to pass data in JSON, and to describe their
JSON structures they have repurposed C struct syntax to describe it.
You can find that draft here:

http://www.ietf.org/id/draft-ietf-alto-protocol-13.txt

While some see no value in schema languages for JSON (or schema
languages in general), there does appear to be an interest in methods
more formal than prose to describe expected content of JSON data. At
the very least, they can provide clarity for protocol specifications.
I'll note that while I was working up one of the examples for my
draft, I found an error in the prose of another JSON protocol on which
I had been working.

-andy