Pieces and parts

Hi all,

In thinking through the recent discussion around the protocol definition, I
think we have a few key pieces that could/should be defined separately.

These are (as I understand things):

0. a data format
   - check! ;)

1. a feed format
   - for more than one annotation
   - we MAY have this in the form of the container docs from the protocol

2. a discovery mechanism
   - "I'm at this URL, are there annotations (on this server, other
servers, etc)?"

3. a notification system
  - "I made an annotation on this URL, how do I let it (it's author,
publisher, CMS, etc) know that I did that."

4. a publishing system--which is mostly (afaik) what's being defined in the
protocol spec
  - "I have annotations, and I want to write them into an annotation system"

These map pretty well to the diagram Doug made that's presented on the main
"promotional" landing page for the Annotation WG:
http://www.w3.org/annotation/

If that list (or one like it) makes sense to everyone, I think it would be
prudent for us to begin collecting information (on the wiki, presumably)
around each of those things and begin spec-ing or explaining how they might
(should or could) be used for annotation.

My sense is that we've got #0 and #4 in some state of completion or
progression, but they're floating heads without the rest of these
components and stories. I.e. we've got enough spec'ed to make loverly
matching silos, but not enough spec'ed to make them work together across
the Web.

Thoughts? :)

Thanks!
Benjamin
--
Developer Advocate
http://hypothes.is/