Re: ACTION-983: same-document reference

I forgot to talk about the note in section 4.2.7:

Note that an empty href attribute indicates a "same document reference", 
whereas the URI from which the document was retrieved does not. Need 
confirmation or otherwise of this under Francois's ACTION-983
This is not true. Same reference documents should not be retrieved either.

I would remove the note and amend the section to say:
If the response is an HTML response and it contains a <link 
rel="alternate" media="handheld" /> element, a proxy should request and 
process the referenced resource, unless the reference is a "Same 
Reference Document" as defined in RFC3986 section 4.4 [REF], e.g. an 
empty href attribute or the URI of the HTML response.


Francois Daoust wrote:
> Hi,
> Discussion on "same-document" references started a long time ago when 
> Dom managed to have the group follow his unwise principle that a URI 
> always represents the resource and not a given representation of the 
> resource. This led to the production of a very smart algorithm in the 
> last call version of the guidelines. This was shortly followed by last 
> call comment LC-2009 [1]. The comment pointed us to section 4.4 of 
> RFC3986 "Uniform Resource Identifier (URI): Generic Syntax" [2] that 
> defines the concept of "same-document reference".
> In particular, it does say:
> [[
>    When a same-document reference is dereferenced for a retrieval
>    action, the target of that reference is defined to be within the same
>    entity (representation, document, or message) as the reference;
>    therefore, a dereference should not result in a new retrieval action.
> ]]
> ... meaning that a URI that appears in the representation of a resource 
> and that happens to be a same-document reference represents the 
> representation of the resource, and not the resource itself.
> We blamed Dom. We still had extensive discussions on the topic such as 
> in [3], in particular because it also connects with the ("Oh no!", the 
> Lemming says and explodes) ISSUE-222 [4] and the TAG Finding On Linking 
> Alternative Representations To Enable Discovery And Publishing [5]. The 
> thing is the theory does not entirely match practice and most (all?) 
> browsers do not correctly handle the case when you want to use a 
> canonical URI for bookmarking purpose. Plus there is no true way to 
> define a URI as the canonical URI for a set of representations [6].
> Whilst this is true, it is not directly related to the definition of a 
> "same-document reference" and does not change its definition either. In 
> short, unless we have good reasons not to, we should stick to the 
> definition of the above-mentioned RFC, and this is exactly what Appendix 
> G.1.4.2 [7] does.
> However, the first bullet point in section 4.2.9 [8] restricts the 
> possibility of a "Same Document reference" to an empty href attribute. 
> For consistency, the text should rather be:
> [[
>  the content is HTML and contains <link rel="alternate" media="handheld" 
> href="[same-ref]"/> where [same-ref] is a "Same Document reference" as 
> defined in RFC 3986 section 4.4 [REF]. In particular, an empty href 
> attribute is a "Same Document Reference".
> ]]
> Francois.
> [1] 
> [2]
> [3]
> [4]
> [5]
> [6]
> [7] 
> [8] 

Received on Monday, 22 June 2009 16:03:37 UTC