Re: [agenda} Web Annotation WG teleconf Agenda (2014-10-01)

Hi Maxence,

On Wed, Oct 1, 2014 at 8:56 AM, Maxence Guesdon <Maxence.Guesdon@inria.fr>
wrote:

> On Mon, 29 Sep 2014 12:57:35 -0400
> Frederick Hirsch <w3c@fjhirsch.com> wrote:
>
> # On
>
> http://www.slideshare.net/azaroth42/open-annotation-core-data-model-tutorial
> 1. Is there a way to provide the "same body" in different formats ?
> Is it a way to use the "choice" described on slide 3
>

Yes, a Choice is how you would give multiple options for the "same" body.
It could also include multiple languages, or other dimensions of
similarity.  It could even be the exact same representation, available from
different URIs.

2. Am I right if I say that all information relative to an annotation
> is the subgraph starting from the annotation IRI and including all
> blank nodes from here, stopping when encountering a non-blank node
> (including this non-blank node in the annotation subgraph) ?
> In other words, is there a simple way to remove/extract all information
> about an annotation from an existing RDF graph, without having to act
> specifically depending on each predicate ?
>

As with any graph based system, appropriate boundaries within the graph are
difficult.   I'm not sure that there's an easy rule that will capture
everything, but that would be a great point to discuss further.



> # On
>
> http://www.slideshare.net/paolociccarese/open-annotation-specifiers-and-specific-resources-tutorial
>
> 1. On "Specifiers - Time state", keeping copies of resources is given
> as a way to prevent an annotation to refer to something not existing
> any more. But is it practically possible ? A web page is usually not
> only one HTML file, but also one or more style sheets, images, ...
>

For every system to do it separately... probably not. However there are
many web archives, such as the Internet Archive (web.archive.org), that do
this both on demand and by crawling.



> 2. Regarding the HTTP request state, should there be a list of relevant
> header fields ?
>

It's hard to say which headers are important and which aren't.  The content
negotiation headers are clearly important, but who knows what the server is
actually doing with all the rest of them?  So we left it wide open to
include as many or as few headers as desired.



> 3. Regarding style, just an idea: could the browser help out in styling
> annotations so that they are visible ? i.e. display in light color when
> the background is dark, ...
>

That might be possible when the browser controls the colors, eg background
color. But for images or videos, it seems like a harder problem for the
browser to determine what the predominant background color is.  Certainly
compared to being told in a native styling language (CSS) how to style the
annotation.

Hope that helps!

Rob

-- 
Rob Sanderson
Technology Collaboration Facilitator
Digital Library Systems and Services
Stanford, CA 94305