Re: Representation of a secondary resource?


I'm sorry but while I'd probably survive my issue being dropped by the
TAG, I wouldn't be satisfied. I guess that's where I'm supposed to say I
cannot live with the text as it stands.

BTW, the URI you provided gets me to a listing of a directory and in
general I noticed that a few editors' drafts have inconsistent URIs,
dates in heading and "this version" URIs.

But back to the point: In order to minimize the impact of my issue's
resolution on your spec, I can live with changes only to section 3.1
"Using a URI to Access a Resource" or 3.1.1, "Details of retrieving a

Neither of these sections currently mentions the special handling
necessary for URIs with fragment identifiers, nor are they in a context
that would imply they only apply to URIs without fragment IDs.

I disagree this issue is a question of degree. I see that the proposed
formulations below can be viewed as such (which was a subtlety that I
missed previously), but the issue is whether one can or cannot expect to
be able to get a representation of a resource identified with a URI
containing a fragment ID. 

This is too fine a point, though, the problem is that sections 3.1 and
3.1.1 don't even mention it. Let's entertain another proposal, a
paragraph to be added in 3.1, possibly after the first current

        "Note that for a URI with a fragment identifier, the client can
        directly access only the primary resource. For further
        discussion of fragment identifier handling see section 3.2.1."

Is this an important point or a minutia? In light of the recent
discussions involving Patrick Stickler I tend to think it's important
enough to be mentioned in AWWW. I believe my change will do more good
than harm. 8-)

Best regards,


On Thu, 2004-10-28 at 19:27, Stuart Williams wrote:
> Hello Jacek,
> I think that these are questions of degree... wrt to what is the general 
> rule and what is the exception. I have come to the conclusion that 
> Webarch says just about as much as it can on the topic, and I couldn't 
> really call the generality one way or the other. I was hoping that you 
> might see it that way too and that I could peruade you to say that you 
> could live with the text as it is in our current editors draft of 
> Webarch at
> Can you indicate whether or not you can live with what is there? Thanks.
> Cheers
> Stuart
> --
> Jacek Kopecky wrote:
> >Stuart,
> >
> >thanks for writing it up. I just want the TAG to fill what I perceive as
> >a hole in AWWW regarding the representations of resources identified
> >with fragment identifiers.
> >
> >Your observation is acute but I believe it's an exception to the text
> >I'd include. In effect, maybe extending (and perhaps softening) of the
> >text to be included in 2.6 would do:
> >
> >        "In general there is no direct way to retrieve a representation
> >        of a secondary resource using a URI with fragment ID, but in
> >        some cases a process may be available for producing the
> >        representation of the secondary resource from a representation
> >        of the primary resource, which would be specified in the
> >        relevant media type specification; see 3.2.1."
> >
> >As for those internal references, I'd only add reference from 3.1.1 to
> >3.2.1 and note in 3.1.1 that it doesn't in any way illustrate a
> >situation where the URI of the resource contains a fragment ID. Backward
> >reference to 2.6 would IMO be unnecessary.
> >
> >Further, this all would be greatly helped by an example, for example
> >having an XML (application/xml) document at
> >
> >
> ><people>
> >  <person id="john">
> >    <name>John Doe</name>
> >    ...
> >  </person>
> >  <person id="jane">
> >    <name>Jane Smith</name>
> >    ...
> >  </person>
> ></people>
> >
> >The application/xml serialization of the element with the ID "john"
> >together with its content is the representation of the resource with the
> >URI
> >
> >Or perhaps in HTML the same representation but with a different starting
> >viewpoint is the representation of the secondary resource.
> >
> >These examples may well be contentious though, so you may ignore them
> >happily. 8-)
> >
> >Best regards,
> >
> >Jacek
> >
> >On Wed, 2004-10-27 at 15:52, Williams, Stuart (HP Labs, Bristol) wrote:
> >  
> >
> >>Hello Dan,
> >>
> >>I have been chatting with Jacek, I guess the net of it is that there is
> >>still something that he wants/needs us to do to satisfy his comment.
> >>
> >>As I understand it Jacek's principle concern is that we very clearly set
> >>expectations about whether it is possible to retrieve directly
> >>representations of a resource that is secondary with respect to a given
> >>URI. I think that he has 'grok'ed the secondary/primary are not classes
> >>of resource but a relation between resources wrt to a single URI.
> >>
> >>I think that Jacek would be satisfied with the inclusion in 2.6 of words
> >>to the effect of:
> >>
> >>	"In general it is not possible to directly retrieve a
> >>representation of a secondary resource using a URI with fragmentID."
> >>
> >>He's also requested some internal cross referencing between 3.1.1
> >>(making it clear that the example there does not elaborate on the use of
> >>fragIds) and either or both sections 3.2.1 and 2.6.
> >>
> >>
> >>Personnally I'm mixed about whether we need to say more than we say in
> >>"2.6 Fragment Identifiers": I think:
> >>
> >>	"The secondary resource may be some portion or subset of 
> >>	the primary resource, some view on representations of 
> >>	the primary resource, or some other resource defined or 
> >>	described by those representations."
> >>
> >>provides some scope to construe that in somecases the representation of
> >>a secondary resource is some part of the representation of the primary
> >>resource. This makes me reticent about making the more general statement
> >>Jacek is seeking - because in some cases there is an effective procedure
> >>that would yield a representation of a secondary resource.
> >>
> >>Jacek... we didn't discuss this when we spoke, but I'm wondering if that
> >>observation I've just made above would be sufficient for you to be able
> >>to 'live-with' the current wording (and maybe the additional
> >>cross-referencing.
> >>
> >>Regards
> >>
> >>Stuart
> >>--
> >>

Received on Thursday, 28 October 2004 18:15:27 UTC