Re: Idea: Annotation Feed Format

Hi Benjamin, all,

Lightweight is certainly good :)

A few points to consider (as non-chair, as always)

* We should not base the weight/complexity of LDP on my half-baked
description of it! :)  Any perceived complexity is likely my failure,
rather than the specification's.

* In my experience of trying to use Atom, the required fields and inherited
semantics tend to get in the way :(  For example, if an atom entry is
equivalent to an annotation, then the requirement for a title for each
entry means each annotation must have a title ... which is a requirement we
currently don't have.  That results in a lot of boilerplate content to fill
up the xml with meaningless content to be valid according to the schema.

* If we don't wrap the annotation within the feed, is there much advantage
to the feed format beyond just a list of URIs?  If we do wrap it, then do
we need to map everything into RDF so we can serialize out as JSON-LD or
other serializations?

None of which should be seen as a deterrent to working on this sort of
thing, just some notes from past experience :)  [ in particular:
http://www.openarchives.org/ore/1.0/atom ]

Rob






On Tue, Feb 24, 2015 at 6:24 AM, Benjamin Young <bigbluehat@hypothes.is>
wrote:

> The current LDP-based protocol propose is a full read/write API for
> dealing with Open Annotation documents.
>
> I've begun to wonder if we might not also be served by having a more
> minimal "feed format" similar to (or based on) RSS, Atom, or
> ActivityStreams (etc).
>
> The end result should be a lightweight wrapper around Open Annotation Data
> Model documents that return them as a collection. This could be done today
> with Atom via <atom:link /> references and auto-generated title, updated,
> etc. tags. Lorestore apparently provides these:
> http://austese.net/lorestore/docs.html#atom
>
> Is there interest in having a separate document that defines one or more
> options for expressing an annotation "feed" in any of these existing
> formats? or (if we must) defining a new one?
>
> Thanks for the consideration,
> Benjamin
> --
> Developer Advocate
> http://hypothes.is/
>



-- 
Rob Sanderson
Information Standards Advocate
Digital Library Systems and Services
Stanford, CA 94305