Re: May be a relevant discussion for eBooks: 'zip archives as first class citizens'

Hi Brady,

What, speaking for the Open Annotation group, would be great is if there
was one accepted method for identifying segments of resources, regardless
of whether they're embedded in a zip/EPUB or live on the web.

So to select a rectangular region of an image in an EPUB, a client would
use the media fragment structure in the same way that it would if the image
was on the web.  Ditto for the other fragment specifications.  I do
understand that this can be done in CFI using a range of two spatial
offsets, and also for video with the temporal-spatial offset mechanism...
which leads to my concern that if there is a W3C way of identifying a
resource within a zip file, there are now two equally valid mechanisms that
would both need to be supported.


And to answer the question about text selection, if there isn't a way to
use a fragment (eg for HTML) then we have two methods: a Quotation based
selector, and a Offset based selector.
Details at:
    http://www.openannotation.org/spec/core/specific.html#RangeSelectors

Thanks!

Rob



On Tue, Sep 17, 2013 at 8:59 AM, Brady Duga <duga@google.com> wrote:

> Fragment ids can be used in EPUB CFI as a check mechanism. So, a path can
> (and should) include any ids of elements in the path. These can be used to
> attempt correction of the CFI if the data changes. I am not sure how
> fragment ids would be used generically for annotations, since they rely on
> authors marking all elements with ids. I also don't know how you identify
> ranges of text that way, though perhaps a mechanism exists.
>
>
> On Mon, Sep 16, 2013 at 10:03 PM, Ivan Herman <ivan@w3.org> wrote:
>
>> Robert,
>>
>> I guess, as far as this IG is concerned, we can/should restrict ourselves
>> to eBooks, more precisely ePub. In the case of an ePub, the EPUB CFI
>> specification:
>>
>> http://www.idpf.org/epub/linking/cfi/epub-cfi.html
>>
>> becomes part of the picture... I am not sure it answers to your
>> requirements, though, because I am not sure whether the existing fragment
>> id-s (in HTML or in SVG) can be combined easily...
>>
>> I guess Brady can help us out for this.
>>
>> Ivan
>>
>> On Sep 16, 2013, at 18:34 , Robert Sanderson <azaroth42@gmail.com> wrote:
>>
>> >
>> > Thanks Ivan, Robin!
>> >
>> > Some ideas towards requirements from the annotation perspective:
>> >
>> > * Resource-in-Zip (RiZ) should be able to be annotated given the RiZ
>> URL.
>> >  -- anno hasTarget foo.epub%!/foo.jpg
>> >
>> > * A fragment of the RiZ should be able to be annotated, using existing
>> fragment specs for interoperability
>> >   -- anno hasTarget foo.epub%!/foo.jpg#xywh=100,100,20,20
>> >
>> > * ... without using the media fragment "extension" mechanism so that
>> HTML fragments can be used
>> >  -- anno hasTarget foo.epub%!/foo.html#para1
>> >
>> > * Extension ".zip" cannot be required for the encapsulating zip file
>> >   -- anno hasTarget foo.__epub__%!/foo.jpg
>> >
>> > * ... or nested zips as RiZ, if that is supported:
>> >  -- anno hasTarget foo.epub%!/chapters/chap1.epub%!/foo.jpg
>> >
>> > * Relative URLs should work to allow embedding annotations within the
>> zip file and have them move around correctly without always referring back
>> to the original copy of the epub, which might change or no longer exist.
>> >   -- eg:
>> >
>> > <%!/annotations/anno1.ttl> a oa:Annotation ;
>> >   oa:hasBody <%!/annotations/bodies/comment1.txt> ;
>> >   oa:hasTarget <%!/foo.jpg#xywh=100,100,20,20> ;
>> >   oa:hasMotivation oa:commenting .
>> >
>> > * Clearly we need to make RDF assertions about resources inside zips
>> for any of these to work correctly.
>> >
>> > Is it worth adding as a discussion item for tomorrow's call?
>> >
>> > Thanks!
>> >
>> > Rob
>> >
>> >
>> >
>> > On Wed, Sep 11, 2013 at 8:13 AM, Robin Berjon <robin@w3.org> wrote:
>> > On 11/09/2013 10:41 , Ivan Herman wrote:
>> > There is a discussion going on the whatwg mailing list:
>> >
>> >
>> http://lists.whatwg.org/htdig.cgi/whatwg-whatwg.org/2013-August/040599.html
>> >
>> >  on whether URI-s should be defined to address zip archives on the
>> > Web. There has been some remarks about why this would be interesting
>> > and the arguments seem to forget that, in fact, ePub files are also
>> > zip archives.
>> >
>> > Yes, I'm still mulling over that thread and might jump in at some point.
>> >
>> >
>> > I am not sure this is relevant for this IG, and those of you who may
>> > be closer to the issues might want to have a look. It may be relevant
>> > for epub-reader-in-a-browser type implementations...
>> >
>> > I think that this is a key component in bringing browsers and epub
>> closer to forming one web, a goal that I think it pretty relevant :)
>> >
>> > --
>> > Robin Berjon - http://berjon.com/ - @robinberjon
>> >
>> >
>>
>>
>> ----
>> Ivan Herman, W3C
>> Home: http://www.w3.org/People/Ivan/
>> mobile: +31-641044153
>> FOAF: http://www.ivan-herman.net/foaf.rdf
>>
>>
>>
>>
>>
>>
>