[core] W3C EXI WG review of IETF CoRE WG Internet Drafts

[Carine Bournez <[e-mail]>]

Dear CoRE WG members,

The W3C EXI Working Group has reviewed two CoRE I-Ds:
http://trac.tools.ietf.org/id/draft-doi-core-parameter-option-01.txt
http://trac.tools.ietf.org/id/draft-doi-exi-messaging-requirements-00.txt
Here is a summary of our discussions.

About section 2 of "CoAP Content-Type Parameter Option"
http://trac.tools.ietf.org/id/draft-doi-core-parameter-option-01.txt

We believe that the choice of a content-type parameter is a good one
(as long as it does get implemented effectively by endpoints). However, 
the cleanest way would be still to use a content-encoding mechanism 
rather than a separate content-type indicating that EXI is used. So
instead of:

'Accept: application/example-exi;sv=1.4'

we suggest:

'Accept: application/example;sv=1.4'
'Accept-encoding: exi'


About the second I-D, "EXI Messaging Requirements"
http://trac.tools.ietf.org/id/draft-doi-exi-messaging-requirements-00.txt

For the schema versioning concern mentioned in section 2.1, we interpeted
the text (in the light of the core-parameter-option I-D ) as
"schemaId will be sent as a content-type parameter". 
We want to highlight the fact that in practice the idea of using a
'Major.Minor' version number might not be needed. For minor changes to the
schema, it may be interesting to use the deviation capabilities of the
EXI processors, rather than offering a new schema for the content-type
and deploy the change on a lot of endpoints. It would be easier to 
advise users to change the schema (to be negotiated) only when really 
necessary, i.e. format changes enable measurable performance improvements.


For the Efficient XML Interchange Working Group,
-- 
Carine Bournez -+- W3C