David,
we have independently drawn the same conclusions, this seems to be most efficient. Numeric similarity search
needs an efficient approach. Figure 2 in
http://www.orthuber.com/wp1.pdf shows that there are only a few steps that linked data are the first which
allow numeric web search with world wide task sharing. The searchable patterns are
HTTP URIs with feature vectors (sequences of numbers).
Sequences of numbers are the natural way to describe quantifiable objects, e.g.
time as one floating point
number (e.g. seconds since 2000), GPS coordinates as two floating point
numbers, complex measurement results
with more than 100 floating point numbers.
Concerning this suggestions and examples for the best (efficient) syntax are
welcome!
You may also look at
http://www.bentham-open.org/pages/content.php?TOMINFOJ/2008/00000002/00000001/21TOMINFOJ.SGM
which shows an attractive application.
Best
Wolfgang
Address:
Dr. Wolfgang Orthuber, Mathematician, Orthodontist, University Clinic of
Schleswig Holstein, Kiel, Germany
Arnold Heller Str. 16
24105 Kiel
----- Original Message -----
From: "David Booth" <[email protected]>
To: "W. Orthuber" <[email protected]>
Cc: "Hugh Glaser" <[email protected]>; "semantic-web" <[email protected]>;
"Linked Data community"
<[email protected]>
Sent: Monday, May 25, 2009 3:45 AM
Subject: Re: URLs instead of URNs (Was URI lifecycle (Was: Owning URIs))
On Thu, 2009-05-21 at 17:08 +0100, W. Orthuber wrote:
David,
>> In short, although semantic web architecture could be designed to permit
>> unrestricted semantic drift,
>>I think it is a better design -- better
>> serving the semantic web community as a whole -- to adopt an
>> architecture that permits the semantics of each URI to be anchored, by
>> use of a URI declaration.
>Absolutement.
Yes, I think also, URIs should be well defined. Up to now I thought they are,
but your article shows that
URIs (which are not URLs)
have not necessarily an unique definition! Moreover URI should be anchored; the
best would be that they
contain a link to all their
definition and further bindingly associated information.
Why not prefer URIs which are (special "defining") URLs, which contain
a link to a file which contains links to all defining
information (unambiguous
information, in multiple languages if wished)?
So the anchor would be at once accessible and there would be exactly
one location for the decisive information.
Yes, the preferred way to do that is quite well described in "Cool URIs
for the Semantic Web":
http://www.w3.org/TR/cooluris
--
David Booth, Ph.D.
Cleveland Clinic (contractor)
Opinions expressed herein are those of the author and do not necessarily
reflect those of Cleveland Clinic.
