W3C home > Mailing lists > Public > www-rdf-interest@w3.org > October 2004

RE: Some thoughts on effective access to "primary" vs "secondary" resources, consistency of descriptions, and bootstrapping the semantic web...

From: <Patrick.Stickler@nokia.com>
Date: Sat, 9 Oct 2004 11:13:36 +0300
Message-ID: <1E4A0AC134884349A21955574A90A7A50ADD1C@trebe051.ntc.nokia.com>
To: <A.J.Miles@rl.ac.uk>, <public-swbp-wg@w3.org>
Cc: <www-rdf-interest@w3.org>, <connolly@w3.org>



> -----Original Message-----
> From: ext Miles, AJ (Alistair) [mailto:A.J.Miles@rl.ac.uk]
> Sent: 08 October, 2004 18:01
> To: Stickler Patrick (Nokia-TP-MSW/Tampere); public-swbp-wg@w3.org
> Cc: 'www-rdf-interest@w3.org'
> Subject: RE: Some thoughts on effective access to "primary" vs
> "secondary" resources, consistency of descriptions, and bootstrapping
> the semantic web...
> 
> 
> This has come up again recently on public-esw-thes@w3.org also.

It clearly is a pervasive issue.

> Re keeping concepts and documents disjoint, see [1][2].

I think that the issue of truly keeping documents and concepts
disjoint hinges upon whether one presumes that resources denoted
by primary URIs (without fragids) always denote documents.

There is *already* existing, significant, widespread usage which
contradicts such a presumption -- and therefore those who continue
to insist on such a narrow interpretation should accept that the
train has already left the station, it won't return, and get over
it.

It is *not* acceptable to presume that a URI without fragid always
denotes, or even usually denotes, an "information-bearing" resource,
such as a document.

So, in fact, whether you use a hash or slash has nothing to do
with "keeping concepts and documents disjoint". They already are
disjoint, in that *any* URI can denote *any* kind of resource so
there is no inherent semantic relationship between any URI and
any URIref with fragid having that first URI as its base.

In fact, the following is completely valid: 

   http://example.com/foo a ex:Concept .
   http://example.com/foo#bar a ex:Document .

The *real* issue is about accessibility of resources denoted by
primary versus secondary URIs.

--

Regarding DanCs comments the wiki referenced in [2]
http://esw.w3.org/topic/HashURI, I agree that it touches upon
the most central points, including the key problems arising
from using secondary URIs for "first class resources".

However, in the wiki Dan makes the following statements
regarding the use of secondary URIs (with fragids)

[
This works well for a lot of RDF content. It's very natural (especially with rdf:ID) to introduce a new conceptual thing and name it with the same action.

Web retrieval does approximately the right thing. When you ask for foo#bar, your client asks for a representation of foo and then (depending on the content-type of the result) may try to find something named bar in that result. 
]

I have a number of concerns/issues with the way this is stated. 

Firstly, it suggests that using a secondary URI somehow works "better" than
using a primary URI. Yet RDF itself doesn't care one way or another. It is
completely *irrelevant* to RDF whether any resource is denoted by a primary
versus secondary URI. 

The only benefit, and a trivial one at best, is that it saves typing a 
few characters in the RDF/XML serialization, by using rdf:ID and only 
typing the local name rather than having to type the entire URI or define
a few entities. And I (and others) can speak from alot of experience that
this "extra work" is insignificant and certainly does not even begin to
outbalance the problems of using secondary URIs.

Secondly, while it is true that "web retrieval does approximately the
right thing" when trying to indirectly access a secondary resource, it 
implies that the ability to indirectly access a secondary resource 
demonstrates the validity and viability of using secondary URIs to 
denote vocabulary terms; which again, is misleading and misses the
fact that *direct* access of resources is far more efficient, scalable,
and can avail itself of the full richness of the web machinery.

While I find it commendable that key problems with using secondary
resources are identified in this wiki, I see no truly significant,
motivating examples of how choosing hash over slash is *better*
than, and provides more benefit and utility than, and incurs fewer
problems than, choosing slash over hash.

Yes, certain members of the TAG argue the hash position and that
primary resources should be deemed to always denote "documents".
But I've yet to see such a position successfully defended (rather
than simply asserted) by clear and irrefutable examples and evidence
from real-world applications.

The only valid, proven use for secondary URIs that I have ever seen,
is by browsers focusing the presentation of a representation
of the resource denoted by the base, primary URI to the (kind of) 
representation of that secondary resource embedded in the
primary representation. And that is because, historically, there
were no such thing as a "secondary resource", just *anchor points*.
Thus, except where the representation of secondary resources are 
intentionally provided as fragments of a representation of the resource 
denoted by the base, primary URI (and there are limited use cases
where such explicit access dependencies are beneficial), I don't see 
any practical value in the use of secondary URIs (with fragids).

The arguments relating to reduced keystrokes and easier editing
of XML serializations are entirely unmotivating.

<CHALLENGE>
I challenge anyone to present significantly motivating evidence
of the superior benefit and advantage to using secondary URIs
(with fragids) rather than primary URIs (without fragids) to denote
vocabulary terms. 

Such evidence should be grounded in real-world
examples, and reflect direct benefit to real-world applications;
particularly regarding access to representations of resources
denoted by secondary URIs.

Alternately, evidence can be provided how the use of primary
URIs negatively impact applications.

NOTE:

The above noted use case of browser view focus is excluded from
this challenge.

Arguments based on the presumption that primary URIs can/should
only ever denote "information-bearing" resources (documents) do
not constitute any form of evidence of actual benefit, and therefore
are irrelevant to this challenge.
</CHALLENGE>


> Re working around limitations of sURI see [3].

I think that, given the practical, negative, ramifications of using 
secondary URIs (URIrefs with fragids) to denote vocabulary terms,
and the lack of clear non-trivial benefit from doing so, that
it is difficult, if not impossible, to justify not recommending
the use of primary URIs (without fragids) to denote vocabulary terms.

Cheers,

Patrick

> 
> Al.
> 
> [1] 
> http://lists.w3.org/Archives/Public/public-esw-thes/2004Sep/0012.html
> [2] 
> http://lists.w3.org/Archives/Public/public-esw-thes/2004Sep/0032.html
> [3] 
> http://lists.w3.org/Archives/Public/public-esw-thes/2004Sep/0016.html
> 
> ---
> Alistair Miles
> Research Associate
> CCLRC - Rutherford Appleton Laboratory
> Building R1 Room 1.60
> Fermi Avenue
> Chilton
> Didcot
> Oxfordshire OX11 0QX
> United Kingdom
> Email:        a.j.miles@rl.ac.uk
> Tel: +44 (0)1235 445440
> 
> 
> 
> > -----Original Message-----
> > From: public-swbp-wg-request@w3.org
> > [mailto:public-swbp-wg-request@w3.org]On Behalf Of
> > Patrick.Stickler@nokia.com
> > Sent: 08 October 2004 11:46
> > To: public-swbp-wg@w3.org
> > Subject: Some thoughts on effective access to "primary" vs 
> "secondary"
> > resources, consistency of descriptions, and bootstrapping 
> the semantic
> > web...
> > 
> > 
> > 
> > 
> > 
> > I draw the SWBP WG's attention to some comments which I feel
> > are relevant to the WG's activities:
> > 
> > 
> http://lists.w3.org/Archives/Public/www-rdf-interest/2004Oct/0086.html
> > 
> > Regards,
> > 
> > Patrick
> > 
> > 
> 
Received on Saturday, 9 October 2004 08:14:21 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:52:10 GMT