Re: Possible buf of SiRPAC (relative URI resolution for "resource" attribute)

Dear Art.

Thank you for your early replay. (And I'm sorry for this late response, but this is just
after my 4 days vacation.)

I agree with you in that SiRPAC's implementation of handling "ID" attributes is correct,
but I still suspect SiRPAC incorrectly implements relative URI resolution for
"resource" or "about" attributes.

Although I don't fully understand the thread you informed me, I suppose that the
discussion is related to "ID" or "bagID" attributes rather than "resource" attributes
I tried to point out. So, please let me clarify the point again.

In my understanding, "ID" attributes are used to define a fragment of a RDF document.
It is just like "name" attribute of <anchor>s in HTML or "id" attributes of <card>s in WML.
I believe the usage of "ID" as defining an fragment anchor is so natural that [RDF M&S]
doesn't even clearly mention it, just resorting to our common sense. (As in the discussion you
informed me, technically it would not be so easy because RDF even doesn't have an official DTD
or MIME type but instinctively "ID" just defines an fragment anchor name of a document.)

So, current SiRPAC's implementation of just concatenating baseURI and ID is quite natural
and definitely to be correct.

If we refer to the resource defined with ID="hello" in the document
http://www.sec.co.jp/rdf/test.rdf
That resource will be referred to as
http://www.sec.co.jp/rdf/test.rdf#hello.

The problem remains, however, in the case of fragment referring side such as "resource" (or "about")
attributes. I expect that relative URIs in the value of "resource" attributes will be resolved just
like
"href" attributes of <go> element in WML. The resulting absolute URI will not be just a
concatenation of
base URI and relative URI as implemented in SiRPAC but rather it should be resolved just like
href attributes in usual markup languages.

Currently, resource="hello" in the document http://www.sec.co.jp/rdf/test.rdf seems to be resolved
into
http://www.sec.co.jp/rdf/test.rdf#hello just as ID attributes, but my expecting result is
http://www.sec.co.jp/rdf/hello.

Similarity between "resource" attributes in RDF and usual "href" attributes is clear considering the
fact
that both of them refer to another resource, and if URI is written in a relative form the resource
pointed should be relative to the base document.

Regards
Ryo Asai


On Wed, Nov 22, 2000 at 10:54:35AM +0900, Ryo Asai wrote:
>
> Does anyone tell me SiRPAC's implementation is right or not.

FYI - this issue was raised a while ago by someone from the
CC/PP WG:

  http://lists.w3.org/Archives/Public/www-rdf-comments/2000AprJun/0013.html

and given the response from Ralph Swick (co-editor of the M&S spec):

  http://lists.w3.org/Archives/Public/www-rdf-comments/2000AprJun/0014.html

my take is that SiRPAC is following the "expectation of the Working
Group".

Note, however, that Ralph states:

  In fact, the RDF Model and Syntax specification does not tell you
  how to construct a full URI for the resource named by an ID or bagID.
  The resources are addressable at best only locally within the
  same RDF/XML expression.

Given that the construction of a full URI is not specified, it seems
like SiRPAC [and any other system that constructs such a URI] should
be "right" or "not wrong" :-).

Art
---

Received on Monday, 27 November 2000 20:12:07 UTC