W3C home > Mailing lists > Public > semantic-web@w3.org > February 2005

Re: Combined Inverse Functional Properties

From: Henry Story <henry.story@bblfish.net>
Date: Tue, 15 Feb 2005 12:29:05 +0100
Message-Id: <1d1c3d614118eaa4e37c69f016ee21d1@bblfish.net>
To: semantic-web@w3.org

Here is the same idea, but put in terms of deductions.

Assuming that R1 and R2 are CIFP relations on Class C1, the following
would be the type of deduction I would be looking for.

a ----R1----> b
a ----R2----> c
g ----R1----> b
g ----R2----> c
a -rdf:type-> C1
g -rdf:type-> C1
-------------------
a--owl:sameAs--> g


On 11 Feb 2005, at 13:09, Henry Story wrote:

>
> Combined Inverse Functional Properties (CIFP) are one solution to what
> is known as the Multi Column Key problem in databases.
>
> Another solution is the one up on the wiki under the name
> CoordinateProperty [].
>
> I have myself come across this problem and found a nice relation 
> between
> both of these ways of thinking of the problem for my special case [1].
>
> As I am working with the perspective of finding a simple mapping
> between java interfaces and rdf, the solution I found does not seem
> to be general enough. For one it requires an extra relation, which in 
> my
> example [1] was easily available, but in other cases may seem a little
> artificial.
>
> There has been some thought on CIFPs recently, and a paper has even 
> been
> published [2] but it seems to go a lot further than what I was looking 
> for,
> and I am not quite sure about its status.
>
> But it [2] did make me wonder whether it would have been completely
> nonsensical to extend OWL in such a way that one could express that
> a collections of relations were together, inverse functional.
>
> Something like this I suppose:
>
> default:MWBlog
>       a       owl:Class ;
>       rdfs:subClassOf owl:Thing ;
>       rdfs:subClassOf
>               [ a       owl:CombinedRestriction ;
>                 owl:onProperties ( default:service,  default:blogId ) 
> ] .
>
> If this makes sense then I can at least allow my java classes to be so
> annotated, and perhaps express the constraints above in some form of
> rules...
>
> But perhaps by doing this I would be stepping on the tail of some 
> dormant
> dragon?
>
>
> Henry Story
>
> [] http://esw.w3.org/topic/CoordinateProperty
> [1] https://bloged.dev.java.net/servlets/ReadMsg?list=users&msgNo=484
> [2] http://osiris.sund.ac.uk/~cs0tco/
Received on Tuesday, 15 February 2005 11:29:09 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 1 March 2016 07:41:44 UTC