Re: [Moderator Action] Annotation UCs

Ah!

Well… the Annotation WG has worked on this. The Annotation Data Model[1] (which currently in CR), has a section on selectors[2] which introduces a possibility to 'anchor' an annotation in a more flexible and standard manner. The structure is actually extensible if needed. B.t.w., because the notion of selectors may be useful beyond the realm of annotation, the WG decided to publish (later) a separate Note that extracts the standard into a separate document to make it more palatable, and also provides a more flexible fragment ID; see [3] for an editor's draft, there is no official publication yet (planned for later in the autumn).

The question is, of course, whether [1-3] covers what annotations in PWP need. If there are some specific issues on anchoring that would require, e.g., and extension of the selectors, or the final standardization of section 5 in [3], then this is worth adding to this document. But, on the other hand, this may go into too much details.

Is this what you were referring to?

Ivan

[1] https://www.w3.org/TR/2016/CR-annotation-model-20160705/
[2] https://www.w3.org/TR/2016/CR-annotation-model-20160705/#selectors
[3] http://w3c.github.io/web-annotation/selector-note/index-respec.html



> On 22 Aug 2016, at 11:31, Boris Anthony <boris@rebus.foundation> wrote:
> 
> 
>>> 
>>> Bonus thought: persistent reliable annotation display in-situ requires a standard way of “mapping” a document; how to locate points in a document to which various annotation types and behaviors are attached.
>> 
>> I am not sure I understand what you mean…
> 
> When making an annotation on a digital text, the annotation must “know” the precise location in the text it is “placed."
> 
> A simple note is anchored to a point in a text.
> A highlight starts and ends at very precise points in a text.
> etc…
> 
> In my experience so far, schemes and tactics of doing this have been left to reader software implementors; Kindle, Kobo, Readmill, Hypothesis, etc all have their own way of identifying such points. This means that an annotation I make on a text in Kindle is not visible — nevermind in the right place — if I move the document to a Kobo, for example.
> 
> There is a separation here of “software approaches” and the notion of “having a map of the document’s terrain” (based on DOM and text strings, usually, though Kindle seems to use a character count based on the file in question… a weak approach that proves how little they value annotation but I digress).
> One of these should be standardised (I assume the latter…)
> 
> By way of analogy:
> The great explosion in open digital cartography would not have been possible if the latitude longitude system had not been adopted as a standard way of referencing points on a globe.
> 
> How do we standardise a way of referencing points in a document?
> (and what does the resulting map look like? ;)
> 
> This may be out of scope for PWP, perhaps even W3C, but it does need addressing by someone…
> 
> (Again, my hunch is DOM + Text string parsing…)
> 
> 
>>> Is it in this group’s, or Annotation WG’s, purview to broach this topic or is that left to the field to define best practices?
>>> 
> 
> B.


----
Ivan Herman, W3C
Digital Publishing Lead
Home: http://www.w3.org/People/Ivan/
mobile: +31-641044153
ORCID ID: http://orcid.org/0000-0003-0782-2704