On Wed, May 27, 2015 at 9:02 PM, Larry Masinter <masinter(_at_)adobe(_dot_)com>
wrote:
Could the working group be chartered with a requirement to align “Object
Signing and Encryption” so that it uniformly applies to CBOR and JSON?
I think the case has been made that the data model differences between
JSON and CBOR are legitimate and I don’t think the IETF should pick one
over the other as a policy, let the sides fight it out in the market, not
in the decision whether to charter a working group.
But getting each side to recognize the legitimacy of the other and making
OSE more independent of the details of the data model would be really
helpful.
Why not XML and http://www.w3.org/XML/EXI/ too?
Because XML is a document format which makes it a bad choice for a data
format.
The order of elements means absolutely nothing in an object.
<a><b/><c/></a> conveys exactly the same information as <a><c/><b/></a>. If
the order of a sequence of objects matters, then a sequence should be used.
SGML/XML offer a lot of flexibility in expression because this does matter
in documents. Documents are edited by hand and so schema validation is
useful.
Moving data in a Web Service is essentially just an API call or a return
[occasionally a callback]. More flexibility is neither necessary or
desirable.