On Thu, Aug 8, 2013 at 3:58 PM, Carsten Bormann <cabo@xxxxxxx> wrote:
On Jul 30, 2013, at 09:05, Martin Thomson <martin.thomson@xxxxxxxxx> wrote:Indeed.
> What would cause this to be tragic, is if publication of this were
> used to prevent other work in this area from subsequently being
> published.
As Paul and I have repeatedly said, CBOR is not trying to be the final, definitive, binary object representation for all purposes. It was written to some specific objectives, clearly stated in the document. It is being proposed for standards-track because specific ongoing work that works well with these objectives will benefit greatly from being able to reference a common specification. If the objectives for other work are different, that work may benefit from using a different format, existing or newly designed.
I do not expect you to succeed in being the final definitive format.
But if you are not prepared to try then I don't want the result.
You keep talking about your design requirements, but where are the use cases that drive them? When I am discussing a specification I am always referring to illustrative use cases that motivate the requirements. When you are discussing your spec you wave away every objection by saying it isn't a design requirement but I have never seen you give the rationale.
Coming back to Martin's comments about the type system. I think that is what JSON does so well. It has exactly as much type system as is useful and no more. The developers realized that there was no need to distinguish lists and sets because both are going to serialize as lists on the wire.
Website: http://hallambaker.com/