The "U" in "URI", and the Art of Memory

I'm a newcomer to RDF, but have always had a deep interest in metadata 
and "resource description", even before I had the words to describe 
them. This is a fairly philosophical post, for which I apologize, but 
recent discussions seem to be revolving around philosophical rather than 
technical issues.

One interesting point that arises from the murky depths of the "URI for 
language identifiers" thread is the possibility of mis-interpreting the 
"U" in "URI". It is easy to decide that "Uniform" in this context SHOULD 
actually mean something closer to "universal"--uniform in meaning. RFC 
2396 is pretty clear, though, in making sure that we know that "uniform" 
refers to structure rather than meaning: the references to semantics are 
in the context of semantic interpretaions of the SYNTAX of the URI, not 
of the component parts.

An ongoing thought-experiment of mine is applying the concepts of the 
Classical and Renaissance "Art of Memory" to resource description and 
discovery. Essentially, these techniques apply a *connotative* namespace 
to a problem of memory. A classic example (paraphrased very loosely from 
Frances Yates' _The Art of Memory_, which I don't have with me at the 
moment) is that of an ancient lawyer preparing for a trial. He would 
associate each element of the case with a familiar object in a specific 
location in a mnemonic "house" built up over time. Perhaps the woman who 
was killed in the case would be an apple or pomegranate (red for blood, 
fruit=fertility=feminine), for instance. These images as they appeared 
in the mind's eye in the context of the lawyer's idiosyncratic house of 
memory, would be consciously freighted with details and associations 
relevant to the case.

Very little difference between that pomegranate and 
http://www.dccomics.com/#clarkkent, is there? "Imaginary" in that most 
of us are accustomed to seeing URIs as URLs, and *as a URL* it's 
meaningless, but freighted with "meaning" by its context, our 
experience, and our tools.

Bringing it home, now: in an "Art of Memory", any requirement that a 
pomegranate as visualized ALWAYS denotes a pomegranate in truth would 
kill the utility of the system. Every practitioner must be free to draw 
his or her own associations, and even the pomegranate tree itself cannot 
say, "...but that's not what I meant!" I agree that the same is true in 
RDF as relating to URIs. Consensus assumptions about the denotative 
meaning of a URI are fine, and useful, but cannot and should not be 
"enforced".

I'm curious... has any work on the application side been done that 
leverages the possible "mis"interpretation of a URI or namespace in 
order purposefully to produce serendipitous results?

Paul Edson
paul@thither.biz

Received on Tuesday, 1 April 2003 13:06:47 UTC