Re: Typing annotations ?

The tagging motivation is definitely the right one to use here.  The larger
question about how to make assertions, rather than just tagging, has been
an open question for a long time however.

In my opinion, the best way to do it *at the moment* is to have a body with
data in the value, rather than a human readable text string.  Just the same
as the SVG question, really.

So a TextualBody, with a value of (for example) '{"color": "blue", "shape":
"circle"}' and a format of application/json.
Alternatives would be to invent your own shape-tagging and color-tagging
motivations, but this clearly doesn't scale.  The Pund.it approach is good,
with a named graph as the body ... however it gets quite far into the RDF
data model.  In JSON-LD 1.1, there will be a much easier way to embed named
graphs and a future Annotation specification might then adopt that approach.

Hope that helps, or at least stirs discussion :)

Rob



On Thu, Nov 9, 2017 at 1:37 PM, Etienne Fréjaville <efreja@wanadoo.fr>
wrote:

> Dear Web Annotations users,
>
> Thank you very much for the 'light speed' answer on the embeded svg
> annotation.
> It's great that the way of achieving this is so simple...
>
> I have another question regarding typing the annotations.
> It's still with embedded bodies.
>
> I'd like to type annotations that would be understood by the annotation
> software.
>
> Suppose that I want to tag some piece of text (identified with the target)
> with either a specific color or a specific shape.
> As there is no property to assign a custom type to an annotation (or I
> didn't find one) what would be the proper way to do this ?
>
> I thought of using (suppose I have 2 annotations) :
>
>   "body": [
>      {
>        "type": "TextualBody",
>        "value": "COLOR",
>        "purpose": "tagging"
>      },
>      {
>        "type": "TextualBody",
>        "value": "blue",
>        "purpose": "tagging"
>      }
>    ],
>
>   "body": [
>      {
>        "type": "TextualBody",
>        "value": "SHAPE",
>        "purpose": "tagging"
>      },
>      {
>        "type": "TextualBody",
>        "value": "circle",
>        "purpose": "tagging"
>      }
>    ],
>
> The idea behind is also to be able to filter (not show) some annotations
> of a particular type...
>
> PS : I used "purpose":"tagging" as it seemed to me the best kind of
> annotation in that situation, but that's probably not the main point of my
> question.
>
> Thank you in advance.
>
> Best regards.
>
> Etienne.
>



-- 
Rob Sanderson
Semantic Architect
The Getty Trust
Los Angeles, CA 90049