[web-annotation] Request header and language negotiation

r12a has just created a new issue for 
https://github.com/w3c/web-annotation:

== Request header and language negotiation ==
[raised by r12a, not yet discussed by the i18n WG]

4.3.2 Request Header State
https://www.w3.org/TR/2016/WD-annotation-model-20160331/#request-header-state

I'm trying to imagine how this works in the case of language 
negotiation. Say, for example, that the user was making annotations on
https://www.w3.org/International/questions/qa-site-conneg

This page is available in two languages: english and swedish. The page
 they'd actually be annotating may depend on a number of things. 
Possibilities include:

a. their browser preferences are set to request swedish versions, ie. 
Content-Language is set to 'sv', so they are working on the swedish 
page.

b. their browser preferences are still set to swedish, but the browser
 has stored information for that site (eg. via a cookie) to say that 
they prefer to read these pages in English, and so the english version
 is what they were looking at while annotating. Now the 
Content-Language information is incorrect if used to point another 
user to the target.

c. they may have originally gone to the swedish page, then followed a 
link to the english page.  Again, the Content-Language information is 
misleading.

I'm assuming that the implementation would supply the HTTP header 
information automatically, rather than the user having to provide that
 information (?).  If so, how does it determine what specific 
information to capture?  And can it record more than one HTTP request 
header (eg. the swedish pdf, rather than the english html, when all 4 
combinations are available)?

Wouldn't the Content-Location in the receiving header be more useful?

Forgive me if i've misunderstood or failed to understand the HTTP 
headers adequately.

Please view or discuss this issue at 
https://github.com/w3c/web-annotation/issues/220 using your GitHub 
account

Received on Tuesday, 17 May 2016 12:56:17 UTC