Re: [Json] Call for WG Consensus on Whether or Not to Adopt Nomenclature Document(s) in the Charter

[Phillip Hallam-Baker <[e-mail]>]

On Fri, Mar 14, 2014 at 7:27 PM, Tim Bray <[e-mail]> wrote:

> I'm generally opposed to this work item on the basis that nobody has
> provided an example of an existing RFC that would clearly be improved by
> the availability of such a thing.
>

I think RFC 6962 could be improved substantially with some greater
adherence to some structure.

Like many JSON Web services, message data is split between the message
contents and the URL.

More importantly, I would like to kill all the parts of the document that
specify the TLS data format (using a schema) and replace them with JSON
structures. But that isn't going to be possible without a better way to
define structures.

I don't think it is just a JSON issue though. While implementing TRANS I
added support for the TLV format to the parser. I have not yet added it to
any of the encoder or decoder libraries though.

It is pretty clear that the proposers of CT are only interested in
addressing notary services for the certificate application and so they are
obsessed with size and compactness. Which is fine. But I certainly don't
wish to carry that data encoding into any other notary application. So
being able to write a schema that allows the data structures to be rendered
in the legacy TLS encoding and in JSON becomes quite important.


-- 
Website: http://hallambaker.com/