Re: [web-annotation] Make Selectors available for the wide world?

@azaroth42 :
> Could you give an example of how just a Selector, with no relation 
to another resource that it selects some segment of, would be useful?

To come up with a very ad-hoc example in RDF, I could imagine a 
foaf-like information that looks something like:

```
<http://www.ivan-herman.net/foaf> a foaf:PersonalProfileDocument ;
    dc:creator "Ivan Herman" ;
    rdfs:seeAlso [
        ex:source <http://www.example.org/teamDescription>;
        ex:refines [
          a selector:FragmentSelector;
          rdf:value "xpointer(/body/section[2]/para[1])"
        ]
    ].
```

where the target of the ``seeAlso`` is a *portion* of the target 
document, and this vocabulary uses the (ad-hoc for now) ``source`` and
 ``refines`` terms. There is no reason to bring a ``SpecificResource``
 into the picture, which is specific to Annotations; any RDF resource 
suffices. 

Of course, the example above could be expressed by

```
<http://www.ivan-herman.net/foaf> a foaf:PersonalProfileDocument ;
    dc:creator "Ivan Herman" ;
    rdfs:seeAlso ex:source 
<http://www.example.org/teamDescription#xpointer(/doc/body/section[2]/para[1])>;
```

because it so happens that ``xpointer`` is a fragment id in XML land, 
but it is not, in fact, in HTML5 land, so it would not be correct 
(that simplification is why I want to have fragment identifiers, too, 
but we agreed about the complications).

Such situations where, essentially, the pure fragment identifier is 
not rich enough to express a selection is, believe, numerous, and 
there is no real solution out there to properly do that. More exactly,
 this group (and, to be fair, mainly the Annotation CG) have come up 
with a viable solution, and the community at large should profit from 
it. 

Note that I do not say we should define *how* selectors should be used
 in general; that would be probably a step too far. This should be 
left to other vocabularies that decide to use selectors.

Editorially, this only means to separate (in the JSON version of the 
model) the current (and revised) section 4.2, with a forward reference
 (i.e., in the main document 'the value of the "selector" term is a 
Selector, see this and this document'). In the RDF document is simply 
means using a different namespace (and make clear in a note that the 
reason of having a selector is to use it in general). I must admit I 
do not see why this is considered to be so utterly complicated:-(


> If Selectors should be split, then why keep State in the Annotation 
spec? Surely being able to describe the representation is a necessary 
first step to selecting a segment of that representation?

I do not have an opinion on that; but I am not sure why the state 
would be the necessary first step as you describe. I do not see this 
strong connection at all (as my example above shows).



-- 
GitHub Notification of comment by iherman
Please view or discuss this issue at 
https://github.com/w3c/web-annotation/issues/110#issuecomment-159569922
 using your GitHub account

Received on Wednesday, 25 November 2015 10:51:54 UTC