Re: The case against URNs

Champion, Mike wrote:
> 
>...

>><p>The <a href="http://www.w3.org">W3C</a> is an important <a 
>>href="http://www.dictionary.com/search?q=organization&r=67">or
>>ganization</a> and <a href="http://www.google.com">Google</a> is an 
>>important service. 
>></p>
> 
> 
> I wouldn't, but I'm not a machine :-) A machine that doesn't understand the
> English meaning of the sentence can't know if it refers to the W3C itself or
> the W3C website. 

That's why RDF exists. So that we can construct sentences that machines 
understand!

Machines don't know anything until we tell them things ANYHOW. If you 
tell them it represents a building and I tell them it represents a 
document then they get confused. But how does "urn:" or "now:" help? If 
you tell them a "now:" URI is a cat and I tell them it is a dog they are 
just as confused!

Plus, we can use RDF assertions to implement the model that Micah 
describes. Existing web pages can represent documents. They can point 
(using RDF) to other HTTP resources that represent concepts (described 
in RDF). You have the same benefits of the pairing of documents and 
concepts, except they live in the same namespace and you can merge them 
into one resource in those cases where it isn't confusing.

So yes, for any particular URI, we should all get together and agree on 
what it means (to whatever extent that is possible). But not for a 
particular *syntax* of URIs. As Jonathon Borden points out, the 
computers don't care about the syntax of URIs unless we tell them to. 
And in lots of contexts, Tim B-L advocates URI opacity himself.

Shhh. If you don't tell computers that using an HTTP URI for a dog is 
confusing, and I don't tell them, then they won't know. It really is 
only confusing to human beings.

  Paul Prescod

Received on Sunday, 6 October 2002 21:45:30 UTC