Re: [Locators] Minutes of meeting 2016-02-10

Thoughts:

With Hypothes.is annotations, when a content URL changes (e.g.
different domain name, and/or moved path) it is possible to specify
the "original" canonical URL using the link element:
<link rel="canonical"
href="https://original.domain.com/original/path/to/content" />
See:
"Hypothesis will coalesce annotations on both pages"
https://hypothes.is/faq/

So, using the Hypothesis HTTP feed API to fetch annotations based on a
URI parameter, this is how data bound to a particular PWP content
locator would be queried (both requests would return the same data
set):

Example 1)
Original canonical URL for a given HTML chapter:
https://hypothes.is/stream.atom?uri=https://domain.com/PWP1/path/to/chapter.html

Example 2)
The chapter URL has changed, but link@rel="canonical" in the
newly-located chapter.html references the original URL:
https://hypothes.is/stream.atom?uri=https://other.domain.com/parent/folder/PWP1/another/relative/path/to/the/same/chapter.html

Now, the above example demonstrates how the resource path itself has
moved (under the PWP1 "folder"). In reality, assuming a PWP transition
from online to offline (effectively, from one origin to another) does
not alter the file structure (i.e. only changes the base URL on which
the file structure is rooted) ; and regardless of whether the PWP
container is zipped or exploded ; then in theory only the central PWP
"manifest" would need to contain the link@rel="canonical" metadata
(much like the base URL of XML and HTML documents, if we consider the
set of linked resources as a whole). The relative path
"./pwp1/path/to/chapter.html" would remain the same, only the anchor
would change from "https://domain.com/" to
"https://other.domain.com/parent/folder/".

In practice though, publishers (or the PWP reading system) may still
need to inject link@rel="canonical" in every single content document,
in order for annotation tools (like Hypothesis) to directly resolve
the correct URI. A specialized, PWP-aware annotation system would of
course be able to process the "manifest" level of indirection, in
order to infer individual document-level canonical URLs from the
combination of the base canonical URL (declared in the PWP manifest),
and the relative path of the targeted resource.

/Daniel

On Thu, Feb 11, 2016 at 9:23 AM, Ivan Herman <ivan@w3.org> wrote:
> Meeting minutes are here:
>
> https://www.w3.org/2016/02/10-dpub-loc-minutes.html
>
> Cheers
>
> Ivan
>
> ----
> 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
>
>
>
>

Received on Thursday, 11 February 2016 10:47:44 UTC