W3C home > Mailing lists > Public > uri@w3.org > August 2001

Re: draft-masinter-dated-uri-00.txt

From: Al Gilman <asgilman@iamdigex.net>
Date: Sun, 26 Aug 2001 11:40:28 -0400
Message-Id: <Version.32.20010826110947.04164c90@pop.iamdigex.net>
To: "Sean B. Palmer" <sean@mysterylights.com>, "Larry Masinter - LMM@acm.org" <lmnet@attglobal.net>, "Aaron Swartz" <aswartz@upclink.com>
Cc: <uri@w3.org>
At 10:38 AM 2001-08-26 , Sean B. Palmer wrote:
>> The tdb form would certainly appear to be useful in
>> creating the sort of URIs one needs for the "predicate"
>> URI in an RDF (subject, predicate, object) triple.  It
>> fits the 'test case' reference in EARL to a T, [...]
>
>And I had to redraft that particular part of the EARL schema in order to
>cope with the fact that people were bound to try to embed the contextual
>information (date or version) within an identifier itself. I simply defined
>a sub class of resources which are date and version-less :-)
>
>Note that earl:TestCase is generally an Object/Class, not a
>Predicate/Property value. Properties can be Objects too, of course.

AG:: No.  The property is intension-equivalent to the assertion that that
object plays the indicated role with regard to the object bearing the
property.  The object is not the property, it is the 'value' of the property.

Properties are memes, patterns, and not objects.  There is an actual 'thing'
which is the uttered articulation of an acceptor.  To cite it with 'tdb' and
you have a system that works.

Yes, we want to make memes "first class" but they are not "first class
_objects_."  By the time you assume they are objects, you have lost the
ballgame.  The difference is of the essence.

Everything should be as simple as possible, and no simpler.  The duality of
the
Boolean algebra is a very strong automorphism.  But the two operators, or the
two morphs, absolutely are distinct.  To make objects of memes is to attempt
the one-point compactification of the Boolean algebra by saying that the
universal set and the empty set are the same.  This is not effective software
engineering, IMHO.


>[...]
>> If this violates a semantic assumption of URNs,
>
>The semantic assumptions for URNs are the same as those for any URI.

AG:: Yes, that's an assumption, but not a fact.  

It fails to be true of the large category of de_facto URIs known as search
URLs.
The service of searching persists.  but there is no persistent association to
anything beyond this verb.  In the case of this class of URIs, one has to
spell
the service request verb-object, and not object-verb, because the identity of
the object(s) recovered is the value-added innovation, what is produced
that is
novel relative to what is consumed, by this service.

>Persistence within a given context. The "urn:" parts simply adds an
>informative "organizational commitment to persistent", which may or may not
>be helpful to humans, and is never helpful to machines.
>
>> then the tdb form should be a new root scheme which
>> may take an URL or URN (encoded as required)
>> including a duri as its 'by' argument.
>
>It could be a new root URI scheme, in that any URN scheme can be a URI
>scheme and vice versa.
>
>Consider the ISBN scheme. Roy Fielding argued that because ISBN schemes are
>reused, they cannot be URNs, but can be URIs. Of course, they *can* be URNs
>because the persistence of an ISBN identifer is good enough to be used...
>an ISBN identifier persistently identifies the book that most commonly has
>that identifier printed on it somewhere. It is the identifier of a
>currently published book using that identifier.
>
>Similarly, URLs can be looked upon as URNs... 

AG:: sure.  Notably that's exactly what news: URLs citing articles by
Message-ID are.  And vice versa.  Names encoded as ISBNs help you locate books
on the shelves and in interlibrary loan.  That is a good bit of their
utility. 
The semantics of URLs and URNs runs together in a way that assuming a
partition
is not a very strong play.  See for historical amusement

 uri@bunyip.com from October 1997: Re: The UR* scheme registry, Citing URL/URI
specs
 URL:
<http://lists.w3.org/Archives/Public/uri/1997Oct/0006.html>http://lists.w3.o
rg/Archives/Public/uri/1997Oct/0006.html

Al

>the URL
><http://www.w3.org/index>http://www.w3.org/index persistently identifes the
resource "index" on the
><http://www.w3.org/>www.w3.org server. The persistence of an identifer always
depends upon the
>scope in which it is being used, and if you go to a very abstract level,
>then you get persistence such that the identifier would qualify as a URN.
>But URN doesn't give a quantitative figure for that, and IMO that's broken.
>
>--
>Kindest Regards,
>Sean B. Palmer
>@prefix : <<http://webns.net/roughterms/>http://webns.net/roughterms/> .
>:Sean :hasHomepage <<http://purl.org/net/sbp/>http://purl.org/net/sbp/> .
>  
Received on Sunday, 26 August 2001 11:20:21 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Thursday, 13 January 2011 12:15:29 GMT