W3C home > Mailing lists > Public > www-rdf-interest@w3.org > November 2001

Re: naming properties and classes in RDF

From: Libby Miller <Libby.Miller@bristol.ac.uk>
Date: Fri, 16 Nov 2001 17:37:28 +0000 (GMT)
To: "Sean B. Palmer" <sean@mysterylights.com>
cc: Libby Miller <Libby.Miller@bristol.ac.uk>, www-rdf-interest <www-rdf-interest@w3.org>, "neil.jacobs" <neil.jacobs@bristol.ac.uk>, "g.conole" <g.conole@bristol.ac.uk>
Message-ID: <Pine.GSO.4.21.0111161724370.1481-100000@mail.ilrt.bris.ac.uk>

> 
> Oh, heh: I get it now. This is a bit concerning, in that people like
> to name their ranges after the property name. :person and :Person,
> :name and :Name. We had quite a discussion about this on SWAG a while
> ago, when we were disucssing using nouns for properties.

I thought I'd remembered something on SWAG.

> 
> It'd be cool to define a sub property of range that is unambiguous:-
> 
>    :unamRange a daml:UnambiguousProperty;
>       rdfs:subPropertyOf rdfs:range .
> 
> then you could do stuff like:-
> 
>    _:a [ :unamRange foaf:Research_interest ] _:b .
> 

that would be useful.
> 
> I conclude that it doesn't actually matter to have classes named after
> properties or vice versa, as long as everything is properly stated.
> For example, there is a property in EARL ("testSubject") that is named
> after its *domain* rather than its range. No one seems to have had a
> problem with that as of yet :-)
> 

I guess so. It just niggles. I'm trying modeling out on people who
don't kow anything about RDF or about modeling. It's a very interesting
experience, as I keep running into problems with my own perception of
how to do things. I'm using RDFAuthor to allow people to create data
visually, with defined schemas. So I'm getting questions like

- if items don't have uris, where do you get them from if you're not a
programmer? what's the best way in terms of minimising cataloguing
error?

- generally, if you're creating instance data, how is it best to avoid
mistakes and confusion? You need to be able to see what you're doing 
as you add to it, such as being able to obviously distinguish
between instances of different types. 

- how do you explain to people how best they should create RDF instance
data (things like the apparent redundancy in the example I gave)?

- how do you divide up the problem so you can create manageable
qualtities of data at a time that aren't too confusing?

I'm not clear how much RDF data will ever be created by generic RDF
creation tools, so maybe these questions aren't that pertinent, but I
thought I'd share them anyway...

thanks for all the comments

libby
Received on Friday, 16 November 2001 12:39:48 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:51:52 GMT