Weaknesses of Annotea.

Aswell as the Fuzzy Pointers (called XPointers by Annotea folks...) there
are other problems I'm having with Annotea, partially I think this is a
difference in view of annotation itself, and partially it's Annoteas
implementation.

Firstly Annotea implementations don't like there not being a context in
the annotation, I have problems with this for a couple of reasons:

 - It limits us to XML documents (or *ML documents where we can
reasonably write a Fuzzy Pointer for.)
 - Content-negotiation makes things very difficult (even text/html
application/xhtml+xml content-negotiation fails - which is relevant with
Annozilla getting different content in that respect to Snufkin (no idea
what Amaya gets.) - Even more extreme forms of content-negotiation, such
as that done by http://annotest.w3.org/annotations , will cause us even
more problems (an xhtml-html conversion of the fuzzy pointer is probably
reasonably safe.)   [*]
 - It prevents us from making comments about the resource, we have to
make comments about part of the resource.

I see no conceptual problem about annotating a resource such as
<mailto:jim@jibbering.com>  and it's something I may well want to do,
also I'd like to say something about a resource rather than just part of
the resource Saying "this is an interesting project" on
http://www.w3.org/2001/Annotea/ is clearly something about the resource
rather than anything about the representation.

I can see there is an argument that such a thing is clearly in the realms
of RDF itself and isn't necessarily an Annotation, but it's certainly how
I see Annotating documents, and believe it would be silly for me to
invent some new RDF purely to repeat what Annotea already has defined
just under too restricted a sense, and to then convince all the other
implementations to also support my schema.

FillyJonk [**] not having any ability to view HTML documents (yet
supports Annotea) I want to use to Annotate resources, why can't I have a
context of the whole resource?

Next is Annotating Images, here the limitation on *ML documents is the
problem, I want to use SVG to annotate areas of bitmapped images,
http://jibbering.com/svg/AnnotateImage.html shows the sort of thing I
mean (the client portion)  ( http://jibbering.com/rdf/foafweb.html shows
some thoughts on SVG in RDF which may be related.)

Cheers,

Jim.

[*] The first two are strictly problems with XPointer over http rather
than problems with Annotea, but Annotea has chosen to use XPointer.
[**] http://jibbering.com/FillyJonk/

Received on Monday, 18 March 2002 11:37:20 UTC