W3C home > Mailing lists > Public > www-rdf-interest@w3.org > February 2004

Re: "Locally-Significant" Statements

From: Thomas B. Passin <tpassin@comcast.net>
Date: Thu, 19 Feb 2004 08:39:27 -0500
Message-ID: <4034BC8F.4030804@comcast.net>
To: "'www-rdf-interest@w3.org'" <www-rdf-interest@w3.org>

Hamish Harvey wrote:

> On Wednesday 18 Feb 2004 2:31 pm, Thomas B. Passin wrote:
> 
> 
>>This is a worthy problem, but it is not unique to RDF.  It is really a
>>data modeling problem, and the best solution will depend on what you
>>want to achieve by your modeling.  You would have the same set of issues
>>- pretty much - if you were designing a relational database.
> 
> 
>>The answers to these questions will depend on what you want to
>>accomplish or express with the data model.  Notice that RDF is not
>>involved here.  Only when you have the data model under control is it
>>time to translate to RDF.  In fact, if you create a relational database
>>model, you can then pick apart the tables and turn them into triples -
>>that gives you your schema, basically.
>>
> 
> There seems to me to be a slight contradiction here. 
> 
> Idiom exists *within* language. RDF is a language, the relational data model 
> is another. Anyone who has learned a foreign language to any depth will 
> attest that idioms do not move easily between languages.
> 

Yes, and to realize a model in RDF you will inevitably develop some 
idiom for the structures.

> I'd be wary of convincing yourself that you are developing a data model 
> independently of any language for expressing it. You may, if you're new to 
> RDF, avoid allowing RDF to have any influence, but you'll never avoid the 
> languages with which are more deeply ingrained to do so. Better, surely, to 
> acknowledge the influence of language, and try to remain aware of it.
> 

Yes, of course you cannot do the modeling in a vacuum, and you will be 
influenced by knowing that you are heading towards RDF.  But experience 
in relational modeing has shown that it is well to start with a 
conceptual model (that is well normalized), and then adapt it to your 
environment, software sytems, etc.  That is the approach I am 
suggesting.  Naturally you will be thinking in terms of resources that 
will become RDF subjects, and you know that relationships between them 
will become RDF predicates, but that's not really different from 
entity-relationship modeling.

Just think of a typical ER diagram.  It is nearly an RDF graph already, 
just missing the URIs and bnodes.  RDF cannot express the constraints in 
an ER diagram, but OWL can probably pick up most of them.

The emphasis should first be on the data model, not on the RDF, even 
though you know you will end up with RDF.  That is, you have to get 
clear on the basics of what you need to express, for you need to express 
them no matter whether you end up using RDF or something else.


Cheers,

Tom P
Received on Thursday, 19 February 2004 08:38:01 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 5 February 2014 07:14:58 UTC