[
Lists Home |
Date Index |
Thread Index
]
Not a troll, exactly. It is a rehearsal. The TAG has
to work the http range issue. The book I referenced a
week or so ago gave me some ideas about this, particularly
that some of the problems of these definitions may be
treatable given a different logic (ie, quantum vs
boolean logic).
It is precisely that there are multiple spaces and this
introduces ambiguity to terms and elements. Why do we
care? Query on "Michael Jackson" and "suit". For this
example, say you get back a lot of articles about his
legal woes and descriptions of his warddrobe. That is
an example of an ambiguous term. It exists simultaneously
in different semantic spaces. As the book points out,
search engines can use combinations of positive and
negative terms to determine which space is 'meant'.
It is a geometric approach to the problem. The notions
of information spaces suggest that a geometric space
is a realistic approach. Remember, the reader is blind
to the writer.
My math stinks, so I come to the smart people with
this idea.
len
From: Bob Foster [mailto:bob@objfac.com]
Wow, talk about trolling! ;-}
How about we back up to: why do we care? It's those damn namespaces.
Suppose there were no namespaces. Suppose instead of namespaces, there
were only globally unique names.
Because it's simpler to use an authority than to pretend we can
calculate a globally unique name, suppose we use inverted domain names.
That is, for a registered domain "x.y", its globally unique names are
constructed with the prefix "y.x". Then we might have
"com.intergraph.bullard.len" as a globally unique name.
End of story. No namespaces. Nobody ever asks if "com.intergraph" is a
"resource", whatever that is, because it's patently obvious it isn't.
We don't need namespaces and never did. Imagine how many person-hours
would have been saved if we didn't have them. ;-}
Bob Foster
Bullard, Claude L (Len) wrote:
> That's the critical observation for this and many other
> threads that rely on ontological commitment to sustain
> communications.
>
> Would anyone care to compare that to URIs as a unit of
> information:
>
> 1. Is a URI a resource?
>
> 2. If it is a resource, what operations are significant?
>
> 3. Are URIs ever ambiguous?
>
> Yes, I know: the permathread from hell.
>
> len
>
> From: Alessandro Triglia [mailto:sandro@mclink.it]
>
> The writer makes choices, but a reader cannot always tell which of those
> choices (if any) convey some semantics in the intentions of the
writer and
> which do not. In other words, an XML document may contain more
information
> than the writer considers significant, but a given reader may not be
able to
> separate the non-significant part from the significant part.
|
