Re: RDF "data model" (was comments on syntax draft)

Graham Klyne wrote:

> At 02:33 PM 12/14/01 +0000, Dave Beckett wrote:
> 
>>   (Dan)
>>   "The RDF Model Theory ([RDF-MODEL]) is a graph consisting of nodes "
>>
>>   ...the Model Theory itself isn't a graph; it's a theory. So we kind of
>>   want to say 'the RDF model is a', except we don't seem to use the word
>>   'model' that way anymore. Same goes for 'the RDF data model', I think.
>>
>>   "The RDF Model Theory" [RDF-MODEL] defines RDF as a graph..." perhaps?
> 
> 
> I was thinking, just overnight, that we seem to have kind of backed 
> ourselves into a bit of a corner over this "data model" idea, and the 
> fact that "model" has a special meaning for logicians.  I think that the 
> abstract "data model" is a key backbone of RDF, yet we haven't really 
> defined a convenient way to talk about it.


Well, there is a data model section in the Primer (which has been 
somewhat rewritten from what's there, but not posted to the Web site 
yet), and I think this is at least one of the places we intend to talk 
about it.


> 
> I think the "data model" is central in the sense that (a) syntax can be 
> defined as a lexicalization of the data model, and (b) semantics can be 
> defined using the "data model" as it's input language for which 
> denotations are defined.
> 
> I personally would be happy to talk about an "RDF data model", and keep 
> in mind that this uses the term "model" quite differently from its use 
> in "model theory" (and state this explicitly somewhere), but I also see 
> that might be felt to be confusing.  Do we have any other adequately 
> descriptive candidates for what to call this?


As an old "database guy", I'm perfectly happy with the idea of an "RDF 
data model" (and note that that was the way "RDF Model" was used in the 
M&S.  My plan is to also briefly describe the model theory in the Primer 
(in non-logician terms), and say something about the difference between 
these uses of the term "model".  I don't feel it's appropriate to cast 
around too far for another term to substitute for "data model" just 
because "model" is also used in "model theory".  "Data model" is a 
perfectly respectable usage in another (sub)discipline, and people who 
are going to deal with the Web in general, and the Semantic Web in 
particular, had better get used to the idea that a given word may have 
different meanings to people in different communities.  However, we need 
to make the distinction clear.


> 
> Anyway, the main point of my overnight concerns was that, while we 
> currently have syntax and semantics documents as work in progress, there 
> doesn't really seem to be an accessible yet definitive starting point 
> for non-logicians that introduces the "data model" independently of any 
> particular serialization syntax. [** see below]  (In my work on CC/PP, 
> it was grokking the underlying graph model that allowed me to really 
> come to terms with what RDF was providing, so I do think this is an 
> important issue.)
> 
> Meanwhile, for the syntax document, I'll offer some words:
> 
> [[[
> RDF is an information representation format that is based on a graph 
> data model.  The RDF Model Theory [xxx] gives a formal definition and 
> model theoretic semantics for this graph model;  this document defines 
> how the graph model can be represented and serialized using XML [xxx].
> ]]]
> 
> [**]
> Maybe the "primer" is the right place for the data model to be 
> introduced.  My concern with this is that a document titled "primer" is 
> not sufficiently definitive.  But I can see that this could work if the 
> primer, syntax and semantics were carefully integrated.


I think the Primer is certainly at least *one* place for the data model 
to be introduced.  However, that doesn't prevent its being explained in 
other documents as well.  Once we have some reasonable text, it could be 
replicated (or linked to, or expanded on) in other places.


> 
> Maybe a common introductory paragraph in each of the three documents, 
> expanding on text something like that suggested above, would provide the 
> needed level of integration and cross-referencing.  I'd be prepared to 
> take a shot at drafting this is it's felt to be a useful strategy.


I don't know that anything as short as a "paragraph" is going to get the job done.  
Why don't you have a look at what's in the Primer now (or I could send you the 
revision), and see if it's along the lines of what you had in mind?


--Frank

 


-- 
Frank Manola                   The MITRE Corporation
202 Burlington Road, MS A345   Bedford, MA 01730-1420
mailto:fmanola@mitre.org       voice: 781-271-8147   FAX: 781-271-875

Received on Thursday, 3 January 2002 08:52:59 UTC