Bijan Parsia wrote:
On 24 Mar 2009, at 13:01, Kingsley Idehen wrote:
[snip]
Bijan,
Is "Identity" important or not? That's the question here.
But it's not a well-defined question.
But let me make a well-defined version (not meant to capture your
question, but just to be clear example). In XSD, 1.0^^xsd:double is
not identical (is numerically distinct from) 1.0^^xsd:decimal. For
counting successors, identity is critical. For numeric equality (for a
given equality theory) and arithmetic (under that theory) it isn't.
Is granularity important or not? A variation of the statement above.
I don't see how that is a variation of any variant of the question above.
303s are just about "Identity" at the datum level within the context
of the Web when using a particular form of HTTP based URI scheme (the
Slash based URI).
I don't understand that at all.
Should we be able to reference a datum and de-reference a
representation of its description via the Web?
Sometimes? Maybe? I don't know? I'm close to not caring? I don't see
the big impact on engineering, frankly, at a substantial level. It
would depend on the context, the notion of datum, etc. etc.
Again, I prefer operational descriptions here. I'm being perfectly
honest and as charitable as I know how when I say I have no idea what
the "datum level within the context of the Web" *is*. Or, actually,
why I should care about it when building an ontology.
And I am not naive in these matters.
Bijan,
I am not assuming naivety on your part.
I just want to focus on the essence of my comments to Michel and Peter
who have issues with the use of 303 redirection to achieve separation of
datum identity from descriptive representation, when using a particular
URI scheme.
I am simply interested in explaining to them what this is trying to
achieve since it remains a strange point of contention. I use the word
"strange" because I believe that the mechanics of the process are
obscuring the fundamental concept in play: Object Identity.
Kingsley
Cheers,
Bijan.
--
Regards,
Kingsley Idehen Weblog: http://www.openlinksw.com/blog/~kidehen
President & CEO
OpenLink Software Web: http://www.openlinksw.com