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

I think I was mistaken thinking the following ...

Jeremy:
> 4: The current spec is how it is because of an editorial oversight. This
> is an honest-to-goodness error, and should be corrected.
[meaning a 'drafting' error rather than a 'design' or 'review' error]

I now believe that it is probable that when the first versions of the
critical paragraphs were drafted the intent was an exclusive reading.

On a re-examination of the archive I think I have understood the
original motivation.

Summary:
--------

In early WDs the serialization syntax and the abbreviated syntax were
separate, with only the new productions in the abbreviated syntax. These
new productions introduced property attributes; they all permitted
identical attributes; hence the semantics of these attributes were made
uniform accross the three productions (desciption with propAttrs,
typedNode with propAttrs and propertyElt with propAttrs). 
The reification reading of rdf:ID, and early versions of para 214, were
in the description of the serialization syntax (a distinct subsection
from the abbreviated syntax); while the descriptions of all the
attributes in the abbreviated form were largely identical across the
different productions in the abbreviated form.


Sources:
--------

The following draft introduced the abbreviated syntax

http://www.w3.org/RDF/Group/9711/WD-rdf-syntax-971114/

but does not describe what an id attribute on any property element
means.

The first text describing what the id attribute on property elements
mean is found in

http://www.w3.org/RDF/Group/1998/01/WD-rdf-syntax-19980111/

This crucially introduces early versions of all critical paragraphs in
distinct sections of the document.

In its discussion of the serialization syntax (section 3.1)
[[[
The value of the ID attribute, if specified, is the identifier for the
node that represents the reification of the property. 
]]]

In the discussion of the abbreviated syntax we have:
[[[
... else if ID is specified R is a new node whose identifier is the
value of the attribute, ...
]]]

At the same time the group were making major advances in their
understanding of reification and how to serialize it. bagID is
introduced in:

Ralph Swick (Dec 8 1997)
http://lists.w3.org/Archives/Member/w3c-rdf-syntax-wg/1997Dec/0087

(The issue list is:

http://www.w3.org/RDF/Group/Syntax/issuesd.html
)
Prior to bagID the same concept was represented by id. Hence, it remains
plausible (but not probable) that this use of id on propertyElt's with
propAttr's was some sort of left-over from the older notation.

I may still explore what happened when the text for the serialized
syntax and the abbreviated syntax got merged; and when RDF:href got
split into rdf:about on descriptions and rdf:resource on propertyElt's.
The latter removes the regularity of the treatment of *all* attributes
accross the abbreviated syntax productions; and hence makes it difficult
to sustain any non-historical motivation for a non-reifying reading of
rdf:ID.

Jeremy

Received on Friday, 19 October 2001 04:47:52 UTC