[annotator-dev] Request for discussion: non-text annotations?

[Robert Casties]

On 26.11.12 23:21, Randall Leeds wrote:
> On Mon, Nov 26, 2012 at 9:15 AM, Robert Casties <casties at mpiwg-berlin.mpg.de
>>>> http://maphub.github.com/api/#comment_example
>>>
>>> Interesting examples. They seem to be afraid to show their full SVG :-)
>>
>> I take that back, it is complete  :-) (But XML inside RDF, argh. and the
>> units...)
>>
> 
> I would encourage you to ignore the RDF. In this thread, we've already been
> talking about a pseudo-SVG-projected-into-JSON field so it's not like we're
> in a position to judge. ;)

No problem, I'm cool with Openannotation :-) I'm just sometimes having a
hard time imagining to write a client implementing the full spec ;-)

> OA stays "pure", in a sense, by making heavy use of standards, but we
> shouldn't feel bound to their serialization of anything when we know that
> developers are using JSON over the wire. Principle of least surprise
> suggests we should stick near these vocabularies wherever possible as long
> as it doesn't cause a total "WTF" when viewed through the JSON lens.

I see it the same way. OA has different goals. It is a reference
ontology about annotations. In (JSON) implementations we make practical
choices but we should not stray too far in the semantics.

In our annotation store we try to use an internal data model that is
close to OA so we will be able to publish the data as OA compatible RDF
once OA has stabilized ;-)

Cheers
	Robert