RE: X-Values: Typed Data Literals for the Semantic Web and Beyond

> -----Original Message-----
> From: ext Sean B. Palmer [mailto:sean@mysterylights.com]
> Sent: 15 October, 2001 18:58
> To: Stickler Patrick (NRC/Tampere)
> Cc: www-rdf-logic@w3.org; www-rdf-interest@w3.org
> Subject: Re: X-Values: Typed Data Literals for the Semantic Web and
> Beyond
> 
> 
> > Following from all of the recent discussions regarding
> > literals and data types, [...] I humbly submit the attached
> > publication for your consideration and comments.
> 
> The first thing of note is that it ignores the recent URI IG 
> publication,
> by sticking to the classical view of URIs rather than the contemporary
> view:-

No. It does not ignore this publication, and I have read that publication
very carefully, several times. And in fact, my proposal addresses the
first bullet of section 3 of the publication in question specifically.

I will make a point to include a reference to it in my future publications
to make this clear.

My proposal can be read as following the contemporary view, such as it 
is defined, in that the classes URP, URT, and URV may each be viewed, as 
are URL and URN, as "a useful but informal concept". 

I do, however, subscribe to the "classical view" and feel it is an error
to discard the formal distinction between URI classes, and that such
distinctions will become even more important in the future (so it's no
wonder if the wording in my proposal reflects that opinion ;-)

> [[[
> Over time, the importance of this additional level of 
> hierarchy seemed to
> lessen; the view became that an individual scheme does not 
> need to be cast
> into one of a discrete set of URI types such as "URL", "URN", 
> "URC", etc.
> Web-identifer schemes are in general URI schemes; a given URI 
> scheme may
> define subspaces.
> ]]] - http://www.w3.org/TR/uri-clarification/

My understanding of the W3C "clarification" is that the terms URN and URL
are
still valid -- they simply do not serve as formal classifications of URI
schemes.
They *do* however serve as useful, informal, classifications of URI schemes
by which we may ascribe some common behavior or characteristics and conduct
general discussions about types of URI.

What this means, in practice, is that applications do not have to know
what e.g. a URL is, but that all knowledge about a given URI scheme must be
defined in terms of that URI scheme in isolation. One can no longer define
behavioral characteristics for e.g. "URLs" which are inherited by URI
schemes defined as URL schemes. It does *not* mean, IMO, that the concepts
URL or URN are no longer meaningful, or that we cannot, for the sake of 
discussion (and hopefully eventually for the sake of applications)
differentiate 
between characteristics of particular URI schemes in terms of well defined
URI classes.

IMO, the W3C's publication is more a clarification of the differing views of
the issue and not a clarification/resolution of the issue itself -- and may 
very well create more questions than it supposedly aims to answer. Not that
it is not a very useful document. It is. But it doesn't say IMO that we
shouldn't
keep talking about URI classes -- only that applications aren't required to
act in accordance to any defined class membership or understand anything
about any URI scheme not defined for the URI scheme itself. At least that's
my understanding/interpretation of the use of the terms "formal" and
"discrete".

It seems to me that distinctions such as URN, URL, URP, etc. are vital
not just as "useful informal concepts" but as formal classifications of
URI schemes defining discrete parititions which serve to instruct an 
application what the dereferencing characteristics of the URI are. URLs are 
a location. URNs are a name that must be mapped to one or more locations.
URPs are WYSIWIG and do not dereference at all. Etc. etc.

I think the W3C might be shooting itself in the foot with its
"clarification".

I think the contemporary view is throwing something very important away.

I think the original authors of the URI/URL/URN specs had it right, but
since
it's taken so long for us to see support for anything other than URLs, and
since so many folks are (ab)using URLs as URNs, folks think the other
classes 
of URIs must not be valid, or do not need formal distinction. Pity...

(Yeah, I know, I always seem to be the odd guy out in left field... ;-)

> Also, you wrote:-
> 
> [[[
> URNs are similar to URLs in that they are expected to resolve to a web
> resource;
> ]]]
> 
> I think that you should put "URNs are similar to URLs in that they are
> expected to identify a Web resource;" to reduce confusion.

Thanks. Yes. That is a better way to word it. Always my struggle... finding
the right words...   ;-)
 
> As to the URP/URT distinction, I guess that "data:" is a URP 
> scheme then,
> yes? It beats the old "URL" definition that the RFC ascribes to it.

Yes. More specifically, it is a URV scheme.
 
> Overall, I like the draft.

;-)
 
> > It is a working draft of a document that I eventually plan
> > to publish in more polished form, [...]
> 
> You'll be wanting to do in plain text, then.

Yeah. Need to hack some nroff, I guess. Or is there a more modern
way to create internet drafts? (When are they gonna take HTML for 
crying out loud?! What is this, the 90's?! ;-)

Cheers,

Patrick

--
Patrick Stickler                      Phone:  +358 3 356 0209
Senior Research Scientist             Mobile: +358 50 483 9453
Nokia Research Center                 Fax:    +358 7180 35409
Visiokatu 1, 33720 Tampere, Finland   Email:  patrick.stickler@nokia.com

Received on Wednesday, 17 October 2001 08:45:48 UTC