- From: Randall Leeds <randall@bleeds.info>
- Date: Fri, 15 Jul 2016 14:12:12 +0000
- To: KANZAKI Masahide <mkanzaki@gmail.com>, W3C <public-annotation@w3.org>
- Message-ID: <CAAL6JQhdhjNSzrspEpbyG=7k-03FwjAwPVJERL9Jj9xbsaAPiw@mail.gmail.com>
I agree and even thought we did relax this to allow just a single position,
but maybe I never followed up with that idea.
I think it's correct for the moment if you simply specify start and end as
the same value. Remember that the range selector describes a range, so the
position selectors don't need to describe a range in this usage but just a
point.
{
"selector": {
"type": "RangeSelector",
"startSelector": {
"type": "XPathSelector",
"value": "/p[2]",
"refinedBy": {
"type": "TextPositionSelector",
"start": 30,
"end": 30
}
},
"endSelector": {
"type": "XPathSelector",
"value": "/p[3]",
"refinedBy": {
"type": "TextPositionSelector",
"start": 10,
"end": 10
}
}
}
}
It seems a little redundant but it's not wrong, I don't think.
On Fri, Jul 15, 2016, 04:36 KANZAKI Masahide <mkanzaki@gmail.com> wrote:
> Hello,
>
> I'm now trying to map Annotator.js object to Web Annotation, and find
> a bit difficult in using refinedBy/TextPositionSelector with
> RangeSelector.
>
> Now, consider we want to annotate a phrase from char 30 of paragraph 2
> to char 10 of the next paragraph. Annotator's rage object looks like
>
> {
> "start": "/p[2]",
> "startOffset": 30,
> "end": "/p[3]",
> "endOffset": 10
> }
>
> Then, straight mapping the above to Web Annotation target might be
> something like
> {
> "selector": {
> "type": "RangeSelector",
> "startSelector": {
> "type": "XPathSelector",
> "value": "/p[2]",
> "refinedBy": {
> "type": "TextPositionSelector",
> "start": 30
> }
> },
> "endSelector": {
> "type": "XPathSelector",
> "value": "/p[3]",
> "refinedBy": {
> "type": "TextPositionSelector",
> "end": 10
> }
> }
> }
> }
>
> However, current Data Model requires (MUST) TextPositionSelector to
> have both "start" and "end". This is not very good fit for mapping
> from Annotator which does not have "endOffset" on stating point nor
> "startOffset" on ending point.
>
> Do you think this requirement could be relaxed in case of RangeSelector ?
>
> Thank you.
>
> --
> @prefix : <http://www.kanzaki.com/ns/sig#> . <> :from [:name
> "KANZAKI Masahide"; :nick "masaka"; :email "mkanzaki@gmail.com"].
>
>
Received on Friday, 15 July 2016 14:12:50 UTC