Re: Last Call: <draft-bormann-cbor-04.txt> (Concise Binary Object Representation (CBOR)) to Proposed Standard

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 






On Sat, Aug 10, 2013 at 11:37 PM, Larry Masinter <masinter@xxxxxxxxx> wrote:
BCP 70 " Guidelines for the Use of Extensible Markup Language (XML)  within IETF Protocols"
attempted to outline some of the design considerations for data representation using XML.
In 2003, it represented the consensus and also the disagreements about what
was "best current practice" at the time.

Section 3 of BCP 70, "Alternatives", lists some of the alternatives and provides
a comparison.

I think what might be missing is an update to BCP 70 or a companion which more
explicitly compares XML, JSON, CBOR and other alternatives in use in IETF protocols.

If you're interested in this work, perhaps you might review BCP 70 and suggest
updates.

Independently of CBOR, it would be worth revisiting this and writing best practices for using JSON.

One of the areas that XML thought it had handled was extensibility. In retrospect, I don't think extensibility was handled anywhere near as well as people imagined at that time. We had to rewrite the SAML schema despite having the input of one of the XML main authors leading the work. What we thought would be a good extensibility mechanism turned out not to be in practice.

At this point JSON has a reasonable extensibility story but not one that I have seen committed to paper...


I think that we could learn something from XML Schema as well. XML Scema and DTDs are really tools for validating documents and they are useful there because they allow the validator to tell the user which line and column an error occurred.

XML schema can be used as a tool for validating input to a Web Service but much of the effort is counterproductive. Who cares which order the elements are presented? I am not aware of any widely used programming language that would, the order of definition of elements in a structure is only visible at all in LISP like languages.


I see a value of using some sort of tool to define what elements do and do not have specified semantics in a protocol and to describe the structure, i.e. a 'schema' but we need something with a tenth the features of XML Schema and we need to address extensibility.

[Index of Archives]     [IETF Annoucements]     [IETF]     [IP Storage]     [Yosemite News]     [Linux SCTP]     [Linux Newbies]     [Fedora Users]