Re: Issue #rdfms-not-id-and-resource-attr

Dave Beckett <dave.beckett@bristol.ac.uk> wrote:

>>> What I've seen so far is an answer that says the spec doesn't allow it.
>>> I feel that doesn't fully answer the issue raised.
>> 
>> I know that Dave seems to feel differently about this, but I see a simple
>> solution to this and empty property elements:
>> 
>>  - Define an empty property element to be an empty literal
>>  - Define an id and a resource to be the reification of the statement
> 
> We *can* do these things - i.e. changing the RDF/XML syntax but we
> need good reasons.  Neatness is one but I feel it isn't strong enough
> to counterpoint breaking existing parsers; make no mistake - this
> would be a change in the meaning.

Sorry for not being clear, but I see a very good reason for this.

It seems that consensus suggests that:

<rdf:Description>
  <propName />
</rdf:Description

should result in an empty literal for the value of propName. In such a case,
something like:

<rdf:Description>
   <propName rdf:ID="foo" />
</rdf:Description>

could only be understood as assigning foo to the reification of the
statement. Similarly,

<rdf:Description>
   <propName rdf:ID="foo" rdf:resource="bar"/>
</rdf:Description>

would also assign foo the the reification of the statement.

Certainly this does change the meaning, but it changes it to be more
consistent and brings it in line with what (IMO) the majority of RDF authors
expect. I doubt that many (if any) RDF authors understand the current
wackiness and even fewer have built documents on it. I think in this case we
can expend the effort needed to fix this rather large mistake in the syntax.

-- 
[ Aaron Swartz | me@aaronsw.com | http://www.aaronsw.com ]

Received on Wednesday, 23 May 2001 18:07:17 UTC