More relevant to this list, OAC will drive some changes to the Zotero
database as well as code. I've written an overview of OAC's and
Zotero's models and outlined what I think needs to be done[3]. Any
feedback or suggestions would be great. It's expected this will go
into some post-2.0 Zotero release.
If you want to fiddle with the document, the source files are
online[4] too. It uses the Sphinx documentation build system[5].
Cheers,
-Nikolas
[1] http://mith.umd.edu/
[2] http://www.openannotation.org/index.html
[3] http://cloud.github.com/downloads/nikolasco/zotero-oac-doc/ZoteroOAC-20091027.pdf
[4] http://github.com/nikolasco/zotero-oac-doc/
[5] http://sphinx.pocoo.org/
> For OAC annotation sources and targets, I think we only need to support a subset of the possiblities within Zotero.
As sources, Zotero annotations (plain text) and notes (HTML) could be
represented using UUID URNs and using the
Content in RDF vocabulary with some MIME type association added on.
All Zotero normal attachment items can be
supported as targets, using either the attachments URL field or a file
URL, depending on whether the attachment is a
link or stored file.
Wouldn't it make more sense to treat annotations and notes as full-fledged URIs?
This is a general issue that need more work in Zotero as well.
Bruce
On Tue, Oct 27, 2009 at 3:32 PM, Bruce D'Arcus <bda...@gmail.com> wrote:
> Wouldn't it make more sense to treat annotations and notes as full-fledged URIs?
E.g. HTTP resolvable ones.
> This is a general issue that need more work in Zotero as well.
By "this" I mean the identity of Zotero items.
Bruce
Could you make your suggestion more concrete? The reason I suggested
UUID URNs is that they are URIs, via RFC 4122. They can also be easily
generated within Firefox (with nsIUUIDGenerator). I considered using
data URLs (RFC 2397), but then they change identity whenever their
contents change, which I consider a bad thing.
I suppose an HTTP URL could be used, but that would necessitate
hosting all notes and each would require a GET request. If, say,
zotero.org wanted to host these and expose appropriate URLs, then I
suppose that's fine but it's not something I'm comfortable suggesting
since I'm not involved with that service. (I guess WebDAV, and similar
things, are another option, but do we really want to require that just
to share annotations+notes?)
Cheers,
-Nikolas
How so? A URI is first and foremost a global identifier. The HTTP URI
just has the added value that you can (not must) GET stuff with it.
> If, say, zotero.org wanted to host these and expose appropriate URLs, then I
> suppose that's fine but it's not something I'm comfortable suggesting
> since I'm not involved with that service. (I guess WebDAV, and similar
> things, are another option, but do we really want to require that just
> to share annotations+notes?)
It depends ultimately on the purposes of this project. It seems that
if exposing annotations and sharing them across the web is included,
then it might be sensible to identity annotations with resolvable
URIs.
I saw the Annotea work cited. It seems to me that work pretty
explicitly says that annotations should be treated like any other web
document. It also seems to me there are good, practical, reasons this
would be desirable.
Bruce