W3C home > Mailing lists > Public > w3c-rdfcore-wg@w3.org > August 2001

Re: containers test0004 baffles me

From: Jeremy Carroll <jjc@hplb.hpl.hp.com>
Date: Wed, 22 Aug 2001 20:53:53 +0200
To: <w3c-rdfcore-wg@w3.org>
Message-ID: <MABBLGKMPIJFCKFGDBEPEEGFCAAA.jjc@hplb.hpl.hp.com>

Sorry my e-mail on this topic got out before it was meant to.
I insert/correct the references etc

> containers test0004 involves the fourth expansion of the property element
> production rule 6.12.


 [6.12] propertyElt    ::= [ omit ]
                      | '<' propName idRefAttr? bagIdAttr? propAttr* '/>'

> By the decision to drop all the special container rules, this expnasion
then
> applies to rdf:li. By paragraph 232 (which is so horrible its number!)
this
> fourth expansion is different from the others, in that it does not reify,
> but the ID is the ID of the resource.
http://lists.w3.org/Archives/Public/www-archive/2001Jun/att-0021/00-part#232

> in my opinion para 232 is an error, and is addressed in the unresolved
issue
> FFFFFF

No, it isn't. None of the issues would allow us to drop para 232 if we
wanted to. We would need to start a new issue first.

> I did some archaelogy on this one, and as far as I could tell, this
> paragraph dates from the early days before the first working group
resolves
> issues 4, 7 ,11 and they still had a certain amount of confusion about
> referring to a resource and referring to the description of the resource.
I
> think it is a hard call whether to delete paragraph 232 or preserve it for
> the sake of backwards compatibility.

I have probably got those issues numbers wrong, and need to go over the
member archives to get the references. I will do that early september if
there is interest.


Dan Connolly wrote:
> > I guess my implementation will raise a "not supported"
> > exception in that case; I don't see sufficient value
> > to implement this.

> test0005 is also difficult for some parser writers (e.g. me). It is fine
if
> you really are using an XSLT transform to process rdf:li, but if you're
not
> you need some fairly horrid special code for processing a corner case that
> nobody in their right mind would rely on. ARP does it, but I personally am
> very sympathetic to a not supported exception.

Jeremy
Received on Wednesday, 22 August 2001 14:44:17 EDT

This archive was generated by hypermail pre-2.1.9 : Wednesday, 3 September 2003 09:38:47 EDT