W3C home > Mailing lists > Public > w3c-rdfcore-wg@w3.org > February 2004

Re: normativity in the RDF document suite

From: Frank Manola <fmanola@acm.org>
Date: Thu, 05 Feb 2004 16:42:33 -0500
Message-ID: <4022B8C9.9090502@acm.org>
To: Pat Hayes <phayes@ihmc.us>
Cc: Eric Miller <em@w3.org>, w3c-rdfcore-wg@w3.org, Brian_McBride <bwm@hplb.hpl.hp.com>

   My general understanding is that there is a difference between a 
document being normative itself (for what it talks about) and being a 
normative reference for what some other document talks about.  In this 
case, my understanding is that Vocabulary is normative in defining the 
RDF Vocabulary Description Language (and hence the boilerplate is OK), 
and that the decision about removing Vocabulary from the normative 
references in Concepts was because Vocabulary is not normative in 
helping to define the RDF Concepts and Abstract Syntax (vanilla RDF is 
defined without Vocabulary, which just defines a specific ... 
vocabulary), not because Vocabulary isn't itself a normative document. 
(BTW:  what order are the references in Concepts supposed to be in? 
It's certainly not alphabetical).

That, at least, was the general idea behind the way I divided up the 
references in the Primer;  but the notion of "normative reference" in 
the Primer could be considered somewhat questionable I suppose, since 
the Primer isn't normative itself.  Anyway, I recall raising this point 
before (and I agree that this whole business could be confusing).

Bottom line:  my opinion is that the boilerplate should be changed only 
for the Primer, and there should be no normative references to the 
Primer.  Vocabulary should remain normative, even though there may be 
informational references to it.

A couple of further points on this "normative" business:

1.  Section 7.2 of the Primer, in describing Test Cases, says

"The test cases are not a complete specification of RDF, and are not 
intended to take precedence over the normative specification documents."

On thinking about this discussion, I suppose someone could read this as 
saying that Test Cases is *not* normative.  If that's a concern, simply 
changing this to:

"The test cases are not a complete specification of RDF, and are not 
intended to take precedence over the *other* specification documents."

would be a simple fix.

2.  If the WG decides to make the Primer normative, there's a paragraph 
in Section 1 that talks about how the Primer is *not* normative that 
would have to be removed.  There would also have to be some text saying 
that in case of disagreement between what the Primer says and what any 
of the other documents say, believe the other documents.

--Frank

Pat Hayes wrote:

> Eric, Brian suggests that we can sort this out without a telecon.
> 
> 
> The boilerplate in the publication drafts of the docs lists 6 documents 
> and refers to them as a single suite of related documents. The same 
> boilerplate occurs in them all, and it says of each one that it is 
> normative. All of which would be fine, except that the WG a while ago 
> decided that the Primer and Vocabulary documents were less normative 
> than the others, and the citations in all of the documents treat those 
> two as informative references.   (This decision for Vocabulary is 
> recorded at 
> http://lists.w3.org/Archives/Public/w3c-rdfcore-wg/2003Jul/0236 . I 
> can't locate the decision for the Primer, sorry.)
> 
> 
> Maybe this makes sense - there could be a distinction between a document 
> which is normative sui generis, and a normatively-inclined reference 
> from one document to another - but I suspect it will be found confusing 
> to readers, and that everything would be a lot clearer if the assignment 
> of normativity in the document preambles, and those in the references, 
> were brought into line with one another.
> 
> 
> One way to do this would be to have two versions of the boilerplate, one 
> for the 4  normative documents and a different one, which makes to 
> reference to normativity (eg does not say that the English version is 
> the only normative version) for the other two.
> 
> 
> Another way to do it would be to say that all the documents are 
> uniformly normative, with a common boilerplate, and to adjust the 
> references sections so that references to Primer and Vocabulary are all 
> moved from Informative to Normative. However, I think this change might 
> need the approval of the WG, or at least the editors.
> 
> 
> ----
> 
> 
> Actually, now I check the documents, they are not at all consistent 
> about what is a normative reference and what is informative.  Each 
> column of the following table indicates a citation as Informative or 
> Normative by the document named at the start of the row:
> 
> 
> DOCUMENT      Primer   Concepts   Syntax   Semantics  Vocabulary  Test Cases
> 
> Primer          -          N         N          N         N          N
> 
> Concepts        I          -         N          N         I          I
> 
> Syntax          I          N         -          I         I          N
> 
> Semantics       I          N         N          -         I          N
> 
> Vocabulary      I          N         N          N         -          N
> 
> Test Cases      I          N         N          N         I          -
> 
> 
>  From which it appears that the Primer is suitably fawning towards all 
> the other documents, even to Vocabulary; that Syntax and Concepts have a 
> slightly jaundiced views of Semantics and Test Cases respectively:  but 
> that otherwise, they seem to largely agree that Primer and Vocabulary 
> are informative and the others are normative.
> 
> 
> I leave all the decisions to you, but (1) I think we should do SOMETHING 
> to rationalize this, and (2) I'm willing to help in any way that I can.
> 
> 
> My own preference, for what its worth, would be to say that Vocabulary 
> at least should be a normative document, and appropriate changes made to 
> the references in the other documents. I can see no rational reason for 
> it not to be.  The above table should then have I's in the first column 
> and N's or blanks everywhere else, requiring edits to the references 
> sections of Concepts, Syntax, Semantics and Test Cases; and the Primer 
> boilerplate would have the sentences referring to normativity deleted.
> 
> 
> Pat
> 
> 
> -- 
> 
> ---------------------------------------------------------------------
> IHMC       (850)434 8903 or (650)494 3973   home
> 40 South Alcaniz St.       (850)202 4416   office
> Pensacola                 (850)202 4440   fax
> FL 32501                     (850)291 0667    cell
> phayes@ihmc.us       http://www.ihmc.us/users/phayes
> 
Received on Thursday, 5 February 2004 16:41:57 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 9 December 2014 23:04:05 UTC