W3C home > Mailing lists > Public > w3c-rdfcore-wg@w3.org > June 2002

Re: Summary and some analysis: New Semantics Initiative

From: patrick hayes <phayes@ai.uwf.edu>
Date: Fri, 14 Jun 2002 18:44:16 -0500
Message-Id: <p05111a17b9302ac659af@[65.217.30.123]>
To: Frank Manola <fmanola@mitre.org>
Cc: w3c-rdfcore-wg@w3.org

>Brian McBride wrote:
>
>>
>
>snip
>
>
>>
>>I would like to hear the WG's thoughts on:
>>
>>   o to address this problem the solution must be normative; W3C 
>>notes and non-normative appendices don't hack it.

I think that all we have to do is to provide a place where other 
normative solutions can be, as it were, attached. The reserved 
vocabulary idea is intended to be just that.

>
>I'm fundamentally supportive of this new initiative, but as a 
>"process" matter I'd like to hear more on exactly what has to be 
>"normative" in what's being proposed.  I understand that the WG 
>needs a normative definition of the MT.  We have Pat's WD.  As I 
>understand it, the new proposal doesn't really involve changing the 
>model-theoretic definition of RDF that we already have (it involves 
>some slight changes to the document text (the introduction), but not 
>changes to the model theory itself).

To be exact, there are two proposals here. One is the Lbase idea 
requiring the new document(s). That in itself is not a technical 
change to the current MT.  The other proposal is to put the 'reserved 
vocabulary' idea into both the Lbase (its already in the draft 
document) *and* into the existing MT. That is a genuine change to the 
MT - its the old 'unasserted triples' idea in a new and I think 
better form - but that is also urgently needed; I think that if we go 
forward without this, or something very similar, we will have a 
Webont layering train wreck.

Maybe it would be best for the WG to focus on this second 'reserved 
vocabulary' point first, as this is what will remove the current 
log-jam which is giving WebOnt so much grief. I think it is fair to 
say that both Peter and Jonathan feel that this device will be enough 
to remove the current immediate problems; that they both approve of 
its use and would prefer having it to not having it; and that, while 
they both would prefer a more radical solution (though not the same 
more radical solution), that such agreement is not likely to be 
obtained for any more ambitious proposal which anyone has so far 
suggested.

>  In addition, we have two new documents:  (a) a description of a 
>technique for defining semantics of languages;  (b) a description of 
>the semantics of RDF in that language.  Now, in the current MT WD, 
>it seems to me that what we've said is normative is the MT, not the 
>technique Pat used for describing it.  If we publish (b) as also 
>normative, what we're saying is that we assert that the semantics 
>given in (b) are equivalent to those in the other MT document.  Now, 
>that certainly means (assuming we're being responsible about what we 
>say) that we understand (a) well enough to verify that (b) actually 
>does specify equivalent semantics, but does that require that (a) 
>itself be "normative"?  What does it mean for (a) to be "normative"? 
>If the proposal were instead to use FOL to describe the semantics, 
>would we also need a document to say that FOL was "normative"?

I hope not. I mean, we write the documents in English, but we don't 
seem to need to say that English is normative, right? (BTW, the 
proposal does use FOL to write the semantics; Lbase is FOL, in a 
slightly unusual syntax.)

>  (Clearly I'm a little confused, but I hope you get what I'm driving at).

To address your specific worries, I think it would be OK if the Lbase 
stuff were presented simply as another way to describe the (same) 
model theory, rather like an expanded version of section 1.6 of the 
current MT draft. If folk are worried about the equivalence, we could 
say explicitly that in the event of any disparity, the MT document is 
normative. The intent of the Lbase isn't to replace the MT with 
something different, only to provide a more useful way of presenting 
the same content; so if we get something wrong (always a risk when 
working in a rush, I agree) then it's wrong.

Pat

-- 
---------------------------------------------------------------------
IHMC					(850)322 0319   cell
40 South Alcaniz St.			(850)202 4416   office
Pensacola,  FL 32501			(850)202 4440   fax
Received on Friday, 14 June 2002 19:44:16 EDT

This archive was generated by hypermail pre-2.1.9 : Wednesday, 3 September 2003 09:49:18 EDT