XML Base and XPath absolutizing of URIs

Hi Joseph,

As you know there is a wonderful debate in the W3C community now about
whether or not namespace URIs should be absolutized.  I'm not on the xml-uri
list, but have mailed this anyway.  Perhaps you could consider posting this
for me if it doesn't show up there.

The newest XBase spec does not mention XPath in the list of affected
specifications, but it is claimed that XBase is required by XLink, which
relies on XPath.

The XPath Recommendation states that URIs are absolutized, but no mechanism
for specifying the base URL is given.  I need to know as soon as possible
whether an erratum to XPath will be issued to state that XBase will be the
way of doing it.  Alternately, will there be an erratum stating that XPath
does not absolutize URIs?

It is very important to our dsig implementers, whose XPath implementations
seem not to absolutize URIs.

By the way, absolutizing URIs does not serve a useful purpose.  The desire
to absolutize seems to be based on the idea that a relative URI in a
namespace declaration might point to different things as a document moves
from machine to machine, and those different documents might alter the
meaning of the namespace qualified element of the document.

Well, there's no point in doing this.  The document's interpretation cannot
be the subject of serious debate unless it is digitally signed.  Even with
an absolute URI, it cannot be guaranteed that the target remains unchanged.
In other words, if you want to claim that the namespace URI is something
more than a string, that it imparts meaning to the document, then you will
be lost unless you compare not URIs but the contents at the URI.  If the
content changes, then it does not matter that the URI has not.  The only way
to do this is to store, within the source document, a digest of the result
of dereferencing the namespace URI.  Now, please don't flood my email box
with statements about how this violates the namespace recommendation because
I am aware of the fact that the namespace rec says that-- that's the bone of
contention people are having in the first place.  All I'm saying is that
absolutizing the URIs is also a violation of the same rec, which says they
should be treated like strings, but if you do want to violate that rec, then
you should first consider the fact that it does you no good.

Whether specified by absolute or relative URI, the digest value of the
namespace qualifying document can be included in a digital signature (if the
namespace URI truly does point to something meaningful).  Once the signature
is effected, the relative URI can remain as long as it always points to a
document with precisely the same digest value (possibly after
canonicalization or other transforms).

The signature is the only way to guarantee that the meaning of the document
has not been changed because of a relative URI in a namespace declaration,
and it does not require you to absolutize the URI to have this guarantee. By
comparison, if you absolutize URIs, it means that the document will only
have verifiable meaning if one can access the original machine that
absolutized the

John Boyer
Software Development Manager
PureEdge Solutions Inc. (formerly UWI.Com)
Creating Binding E-Commerce
jboyer@PureEdge.com


-----Original Message-----
From: Joseph M. Reagle Jr. [mailto:reagle@w3.org]
Sent: Thursday, June 08, 2000 8:45 AM
To: John Boyer
Subject: XML Base WD re-enter Last Call


Forwarded Text ----
 Date: Thu, 8 Jun 2000 09:27:01 +0200
 From: Daniel Veillard <Daniel.Veillard@w3.org>
 To: chairs@w3.org
 Organization: World Wide Web Consortium (W3C http://www.w3.org/)
 Subject: XML Base WD re-enter Last Call
 Status:  O

 On behalf of the XML Linking Working Group [1], I am pleased to announce
 the publication of a new "XML Base" Last Call Working Draft.

 The document addresses is:

        http://www.w3.org/TR/xmlbase

 This draft was made public 7 June 2000

 The Last Call review period will end 28 June 2000. Please
 send review comments before that date to
 www-xml-linking-comments@w3.org (archives available at [2]).

 XML Base, is a missing piece of the XML infrastructure needed for XLink:

 ----------
 XML Base proposes a syntax for providing the equivalent of HTML BASE
 functionality generically in XML documents by defining an XML attribute
 named xml:base.
 ----------

 The mission to bring XBase to REC status was given by the XML Coordination
 Group 1999-1129 [5] and the XML Linking WG expect it meets the requirement
 of a general purpose base mechanism as defined by RFC 2396 [6]

 At their June 1 2000 teleconference [3], the XML Linking
 Working Group decided to resubmit XML Base WD to Last Call.
 The large amount of comments and the changes made to the
 document led the group to estimate that a second Last Call
 was in order. The disposition of comments document for the
 first Last Call is available at:

   http://www.w3.org/2000/06/xmlbase-comments-20000607.html


 DEPENDENCIES AND REVIEW

 All Group Chairs are invited to send their comments to
 the public list.

 The Working Group has identified dependencies with the following W3C
 Working Groups and requests review from them. Even if review
 is not possible, we request that the Chairs listed below
 announce their review intentions to www-xml-linking-comments@w3.org
 (a public mailing list).

 Here is the list of the groups which we think should should review
 XBase Last Call working drafts, based on our WG charter and
 the history of the XML Linking WG:

     - HTML WG, Steven Pemberton <steven.pemberton@cwi.nl>
     - SVG WG, Chris Lilley <chris@w3.org>
     - I18N WG, Misha Wolf <misha.wolf@reuters.com>
     - DOM WG, Lauren Wood <lauren@sqwest.bc.ca>
     - SMIL WG, Aaron Cohen <aaron.m.cohen@intel.com>
     - XSL WG, Sharon Adler <sca@watson.ibm.com>,
               Steve Zilles <szilles@adobe.com>
     - XML Core, Paul Grosso <paul@arbortext.com>,
                 Arnaud Le Hors <lehors@us.ibm.com>

 There are also the groups from the XML activity which may have some
 feedback since we define the language for XML fragment identifiers and
 this may impact their work:

     - XML Query, Paul Cotton <pcotton@microsoft.com>
     - XML Schema, Michael Sperberg-McQueen<cmsmcq@w3.org>,
                   Dave Hollander <dmh@commerce.net>

 Thank you,

 Daniel Veillard,
 for Eve Maler, Daniel Veillard, XML Linking Working Group co-chairs

 [1] http://www.w3.org/XML/Group/Linking.html
 [2] http://lists.w3.org/Archives/Public/www-xml-linking-comments/
 [3]
http://lists.w3.org/Archives/Member/w3c-xml-linking-wg/2000Jun/0003.html
 [4] http://www.w3.org/XML/Group/1998/09/linking-charter.html
 [5] http://lists.w3.org/Archives/Member/w3c-xml-cg/1999Dec/0006.html
 [6] http://www.ics.uci.edu/pub/ietf/uri/rfc2396.txt.

 --
 Daniel.Veillard@w3.org | W3C, INRIA Rhone-Alpes  | Today's Bookmarks :
 Tel : +33 476 615 257  | 655, avenue de l'Europe | Linux XML libxml WWW
 Fax : +33 476 615 207  | 38330 Montbonnot FRANCE | Gnome rpm2html rpmfind
  http://www.w3.org/People/all#veillard%40w3.org  | RPM badminton Kaffe

End Forwarded Text ----

_________________________________________________________
Joseph Reagle Jr.
W3C Policy Analyst                mailto:reagle@w3.org
IETF/W3C XML-Signature Co-Chair   http://www.w3.org/People/Reagle/

Received on Thursday, 8 June 2000 17:43:07 UTC