Re: URIs and Unique IDs

This discussions reminds me very much of Goodman's grue and bleen. It  
seems to me that one could define some predicates such as:

ex:isBleen a rdfs:Property ;
	rdfs:label "is bleen" ;
	rdfs:comment "If this predicate is used before 1 Jan 2010, something  
which appears blue" ,
		"If this predicate is used after 1 Jan 2010, something which  
appears green" ;
	rdfs:range "True or false"^rdfs:comment .

ex:isGrue a rdfs:Property ;
	rdfs:label "is grue" ;
	rdfs:comment "If this predicate is used before 1 Jan 2010, something  
which appears green" ,
		"If this predicate is used after 1 Jan 2010, something which  
appears blue" ;
	rdfs:range "True or false"^rdfs:comment .

And these can be used safely without any contradictions. Sure, on  
occasion we need to know the time when the predicates were used - but  
that only becomes an issue when we want to translate them out of the  
bleen/grue terminology and into the more usual world of blue/green.

The obvious argument is that the predicates ex:isBlue and ex:isGreen  
would be more useful, because one could use them without recourse to  
time, but in fact, it should be noted that ex:isBlue and ex:isGreen  
can be safely defined as:

ex:isBlue a rdfs:Property ;
	rdfs:label "is blue" ;
	rdfs:comment "If this predicate is used before 1 Jan 2010, something  
which appears bleen" ,
		"If this predicate is used after 1 Jan 2010, something which  
appears grue" ;
	rdfs:range "True or false"^rdfs:comment .

ex:isGreen a rdfs:Property ;
	rdfs:label "is green" ;
	rdfs:comment "If this predicate is used before 1 Jan 2010, something  
which appears grue" ,
		"If this predicate is used after 1 Jan 2010, something which  
appears bleen" ;
	rdfs:range "True or false"^rdfs:comment .

Few, if any, triples are universally true in an ever-changing world.  
foaf:names can change on marriage or by deed poll; colours of items  
can change easily (e.g. leaves going brown in autumn, buildings being  
painted, etc); the meanings of words in the English language evolve  
over time (e.g. the word 'web' would not be understood in the same  
sense as is meant by most people on this list to people a hundred  
years ago).

To fully reason with triples, one needs to understand the temporal,  
geospacial, linguistic, social and personal context in which they  
were generated. However, without that knowledge, an albeit limited  
degree of reasoning is still allowed. e.g. without knowledge of time,  
we can still reason about bleen and grue objects - we just can't  
convert that knowledge into knowledge of green and blue objects.

Something to mull over.

More on grue and bleen:
http://en.wikipedia.org/wiki/Grue_and_Bleen

-- 
Toby A Inkster
<mailto:mail@tobyinkster.co.uk>
<http://tobyinkster.co.uk>

Received on Monday, 1 December 2008 11:28:47 UTC