URCs and structured data


I read your canonical URC paper a few weeks ago, just as the reject notice
came through.  I havent studied the follow on discussion yet (time ...) 

But now I have a clear need driven by the development of an annotation
protocol.  This also relates to the suggested use of URCs for ratings
and whether annotations should be considered metadata.  Leaving all
that aside, the clear need is for a structured data transfer standard,
i.e. a data format.

We need a structured data transfer standard independent of how it might
be used.  URCs could use it, and results returned by the annotation
protocol could use it.  We don't need to think too much about all the
applications of the format to figure out how to make it general enough.
There are several examples of such formats, such as ASN.1 and SGML, but
we need something very simple too, so it will catch on.  Several
standards could coexist, so a simple format doesnt have to satisfy
everyone, just most people.  An extensible format would be great too
but complete generality is not the highest priority.  Usability is
the highest priority.

I really liked the format you had been working on last year that was
based on mail header style, with structured values.  Something along
those lines is what I would like to work on I think.  Another simple
alternative is that used by PDRM, with parens for nesting.  I think
SGML tags and attributes are too confusing for most people, but it
might suffice.  Web browsers are probably close to being able to parse
it which is a big advantage.

I'm going to pick a format for this annotation protocol very soon, so
I'd like to make some speedy progress on it.   What do you think?