W3C home > Mailing lists > Public > public-webarch-comments@w3.org > October to December 2004

Definition of 'namespace document' in glossary still incorrect

From: <Patrick.Stickler@nokia.com>
Date: Tue, 7 Dec 2004 12:07:35 +0200
Message-ID: <1E4A0AC134884349A21955574A90A7A50ADDE1@trebe051.ntc.nokia.com>
To: <public-webarch-comments@w3.org>

Regarding the latest edition of AWWW [1], which attempts to address an error
in the definition for 'namespace document' which I pointed out earlier
to the TAG [2], the new definition is, unfortunately, equally erroneous as 
it continues to impose a particular interpretation on the URI used as 
a namespace name, which is not licensed by the XML Namespaces specification,
nor IMO fully consistent with the principle of URI opacity.

The new text states:

Namespace document
   If an XML Namespace URI identifies an information resource, then that 
   resource is the namespace document for that namespace

However, there is no constraint which prohibits a URI used as a
namespace name to identify an information resource which has no
information nor relationship whatsoever to any terms grounded
in that namespace or any use of those terms.

I.e. users are free to use a URI identifying an information resource
as a namespace name without any intention of treating that information
resource as a namespace document.

True, it may be more useful if the URI used as a namespace name
identified a namespace document, but there is no *requirement*
that it do so, even if it identifies an information resource.

And therefore, one cannot presume that every URI used as a 
namespace name which identifies an information resource also
identifies a namespace document.

One might very well use a URI which identifies the home page
of an organization as the namespace name by which terms of
that organization are partitioned, and presuming that because
the URI identifies an information resource (i.e. a home page)
it also identifies a namespace document is simply erroneous.

In fact, it is fair to say that the great majority of URIs
currently employed as namespace names were/are never intended
nor percieved as identifying namespace documents; and hence,
the present definition imposes an interpretation on existing
usage which was, in many/most cases, never the intent of the
owners of those namespaces, and is thus in immediate conflict 
with existing practice and usage.


I again propose to the TAG that the definition of 'namespace document'
not reflect any presumption about what any given URI used as a
namespace name might identify, but to restrict the definition of
'namespace document' to the distinguishing characteristics of
that class of resource, and at most, to indicate that it is
considered useful to use URIs which identify namespace documents
as namespace names, without erroneously asserting that every URI 
used as a namespace name identifies a namespace document.

I again offer explicit text which I think is more appropriate,
and of course the TAG is free to modify or reject the following
proposed text as it sees fit.

Namespace Document
    An information resource that contains useful information, 
    machine-usable and/or human-usable, about terms in a
    particular XML namespace. It is useful, though not manditory,
    that the URI employed as a namespace name identifies a namespace




Patrick Stickler
Nokia, Finland

[1] http://www.w3.org/2001/tag/2004/webarch-20041203/
[2] http://lists.w3.org/Archives/Public/public-webarch-comments/2004OctDec/0156.html
Received on Tuesday, 7 December 2004 10:08:05 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 18:55:25 UTC