On 11 Aug 2006, at 21:46, Booth, David (HP Software - Boston) wrote:
[snip]
(For example, one probably would not use
a wikipedia URL as a concept identifier in a Semantic Web application,
because the definition of a term can change as fast as a user can type
in a browser.)
[snip]
Actually, I think wikipedia URLs have huge practical strengths as
semantic identifiers (for one thing, they are incredibly easy to look
up via Google), and my sense is that they will, de facto, play a
significant role in many semantic web applications.
The fact that the text of the definition of the terms is editable is
not an obstacle to URI acquiring stable meaning, since usage (i.e. in
this case references to the URI) is what really drives meaning.
Definitions are secondary to usage, when it comes to meaning - when
the two conflict, it tends to be the definitions which change to
reflect usage.
Wikipedia URIs are widely used, and this serves (by context) to
define what they refer to, and to stabilize that meaning - without
preventing it from evolving gradually over time.
A widely used wikipedia URI, such as, say:
http://en.wikipedia.org/wiki/Dog
has a pretty stable meaning.
Yes, someone could edit it at any time to replace it (temporarily)
with a big picture of a cat, but this doesn't actually threaten the
meaning of the URI, anymore than someone hacking into the OED website
and changing the definition of the word 'dog' in the dictionary would
be disastrous for the English language. In both cases, the meaning is
stabilized by usage, and it would recover...
Matt
This email has been scanned by Postini.
For more information please visit http://www.postini.com
