Re: Indirection

[snip]

I think we disagree slightly, but I accept that 
you're right about what Thesaurus authors think their 
data structures mean. 

The problem is probably easiest when we think about nodes in a 
thesaurus graph that 'stand for' individuals. I could have a node
in SKOS thes I make that stands for you. Libby Miller might 
make a quite distinct SKOS thesaurus some years later, and
also include a concept for you. There are then two things there,
both concepts, but there's only one you. SKOS keeps the nodes
separate, as I understand it, so that the node that really 
stands for you (ie. has an rdf:type of Person) carries 
properties such as age, favouriteSong, workplaceHomepage; while
the two SKOS concept nodes have properties that are properties
of conceptualisations of you (eg. dated 2004 or 2007, pointers
to the scheme they come from, etc.).

This is a re-hash of our discussion around the 'denotes', or 
'stands for' or whatever property, I think...

Maybe it can be postponed until after the first WD? 

Dan

Received on Friday, 4 February 2005 18:36:01 UTC