W3C home > Mailing lists > Public > public-annotation@w3.org > August 2015

Re: Annotate same text fragment.

From: Robert Sanderson <azaroth42@gmail.com>
Date: Mon, 3 Aug 2015 16:03:57 -0700
Message-ID: <CABevsUFoNdyNXCdUXdNbJmao9qW=nxhmfxpva_+vZ_tqkXfrew@mail.gmail.com>
To: t-cole3 <t-cole3@illinois.edu>
Cc: Annotation WG <public-annotation@w3.org>
Oh no, the hard questions ...

If I recall correctly, the use case was image segments.

That if an image server was available that could deliver arbitrary segments
of an image, then you would want to retrieve the pixels being commented on
if possible.  Thus the SpecificResource should be the part of the image,
not the RDF description of the part of the image. The Selector, on the
other hand, doesn't have a non-RDF representation so should always just be
the RDF.  In other words, the SpecificResource /is/ a resource that has the
specific representation of the thing that you mean to identify.

However, as these are -protocol- requirements, I think we should pull them
out of the -model- and discuss in terms of our actual protocol.  If the
SpecificResource really has to return the segment (or nothing) then it
would be a NonRdfSource in LDP terms ... and getting from there to its RDF
description would be by following a link header.  That seems more than
mildly inconvenient!

Rob


On Mon, Aug 3, 2015 at 2:42 PM, Timothy Cole <t-cole3@illinois.edu> wrote:

> Rob-
>
>
>
> 4.1 of the Web Annotation Data Model notes
>
>
>
> 1.       If a Specifier (e.g., an oa:Selector) "has an HTTP URI, then its
> description, and only its description, must be returned when the URI is
> dereferenced."
>
> 2.       But if an oa:SpecificResource "has an HTTP URI, then the exact
> segment of the Source resource that it identifies, and only the segment,
> must be returned when the URI is dereferenced."
>
>
>
> I take this to mean that if I am using an oa:SpecificResource to target a
> text fragment as my Annotation Target, I should give it a URI only if I
> have a service that will deliver that exact text fragment when the URI is
> de-referenced. Otherwise I should give a URI to the description (e.g., in
> JSON) of the Selector and then reuse that to construct a new blank node
> oa:SpecificResource in each subsequent Annotation that wants to target the
> same text fragment.
>
>
>
> First, is this the correct interpretation?
>
>
>
> Second, if so this is mildly inconvenient. Can you remind me as to the
> rationale for doing this (e.g., as opposed to allowing a redirection to an
> RDF description of the SpecificResource Target as an option)?
>
>
>
> Thanks,
>
>
>
> Tim Cole
>
>
>
>
>
> *From:* Susanna Netseven [mailto:martinelli@netseven.it]
> *Sent:* Friday, May 29, 2015 3:09 AM
> *To:* Robert Sanderson <azaroth42@gmail.com>
> *Cc:* Annotation WG <public-annotation@w3.org>
> *Subject:* Re: Annotate same text fragment.
>
>
>
> Thanks.
>
> Usually we use Target as oa:SpecificResource cause we need to annotate
> partial text or partial images.
>
> We think this way will permit in future to have informations about
> annotations that share the same text fragment. (eg. SPARQL interrogation)
>
>
>
> Susanna.
>
>
>
> On Thu, May 28, 2015 at 6:42 PM, Robert Sanderson <azaroth42@gmail.com>
> wrote:
>
>
>
> Thanks for the question Susanna.  Yes, that's correct -- you can reuse the
> Target IRIs when they're Specific Resources in the same way that you could
> reuse Target IRIs when they're web pages or other content.
>
>
>
> Rob
>
>
>
>
>
> On Thu, May 28, 2015 at 9:40 AM, Susanna Netseven <martinelli@netseven.it>
> wrote:
>
> Hi, i have a technical question regarding a possible implementation about
> a typical use case we are facing.
>
> I annotate a text fragment in html page, for example by commenting a text
> fragment with "this is incorrect"
>
> In the RDF model we would  have:
>
>
>
> <Annotation-IRI1> <oa:hasTarget> <*Target-IRI1*>
>
>
>
> and other target's predicates for selector for text-fragment, source and
> scope .
>
>
>
> Then, another user wants to annotate the same text fragment.
> in this case the correct RDF should be as following?
>
> <Annotation-IRI2> <oa:hasTarget> <*Target-IRI1*>
>
> In this case we don't assume to create a new Target IRI but reuse the
> first.
> Could be this a correct solution?
>
> Thanks.
>
>
>
>
>
> --
>
> Rob Sanderson
>
> Information Standards Advocate
>
> Digital Library Systems and Services
>
> Stanford, CA 94305
>
>
>



-- 
Rob Sanderson
Information Standards Advocate
Digital Library Systems and Services
Stanford, CA 94305
Received on Monday, 3 August 2015 23:04:25 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 18:54:39 UTC