Re: Revisiting Authoritative Metadata (was: The failure of Appendix C as a transition technique)

On Fri, Feb 22, 2013 at 10:22 PM, Mark Baker <> wrote:


I read this document, which specifies TAG's findings on "Authoritative
Metadata". As a personal opinion, I agree to these findings. Lot of web
implementation already exists, which relies on these concepts (for e.g.,
Content-Type header field within HTTP and MIME).

Its a good web architectural principle, to have a metadata that introduces
the representation. This makes the web sender and receiver systems more
efficient wrt processing data that is exchanged. If inferring the sender's
functional intent only from the data itself, was the only option available
that would make data processing less efficient (in the worst case, this
would require reading the whole data stream to infer the functionality
represented by the content).

Where this metadata should be kept (for e.g as external reference,
encapsulating or embedded) are valid design choices wrt this question. The
table here, shows examples
how different kinds of data systems, would have authoritative metadata
placed where (I've no issues agreeing to these findings).

Having said all this, I'll find it still interesting to have TAG revisit
this topic and issuing any changes in this design area.

Mukul Gandhi

Received on Saturday, 2 March 2013 03:36:33 UTC