This is an archived snapshot of W3C's public bugzilla bug tracker, decommissioned in April 2019. Please see the home page for more details.
Xpath2 (and Xquery) says An implementation's choice to support the [XML 1.0] and [XML Names], or [XML 1.1] and [XML Names 1.1] The [XML 1.0] links to the references appendix which in turn links to the undated versions of the XML rec URI, so links to the "latest" version of XML 1.0. If the currently proposed 5th edition of XML 1.0 is adopted, the Name production in XML 1.0 will change to match that of XML 1.1. If that happens it would appear to make an XPath processor that enforces the current XML 1.0 rules non conformant. Is that the intention, or should the references section be changed to refer to (say) the 4th edition using the fixed, dated URI.
If the proposed revision of XML 1.0 is accepted then I think we have to revisit our whole policy of referring to undated versions of other specs. This is a policy that relies on trust: we are trusting the authorities responsible for those specifications to keep them extremely stable. If that trust is no longer well-founded, such that a change in those specs can make existing implementations non-conformant, then we have to move back to referring to dated versions.
This is a valid question. The answer depends to some extent on what the XML Core Working Group decides, so I'm accepting the issue without attempting to resolve it yet.
(In reply to comment #1) If that trust is no longer > well-founded, such that a change in those specs can make existing > implementations non-conformant, then we have to move back to referring to dated > versions. > Actually on further investigation, I don't believe that referencing dated specs will solve the problem. even if you refer to the 4th (or earlier) edition explicitly then its standard header section says that the there may be normative corrections in the errata, and this has _already_ been done. The 4th edition errata http://www.w3.org/XML/xml-V10-4e-errata Already has published E09 which changes the Name production, so this is already the offical XML syntax whether or not the 5th edition is published. By making this change via the errata system, basically the intention of the XML core WG is to remove choice and force all XML systems to switch (which is explicitly the intent, following the lack of take-up of XML 1.1). David
The status section of the 4e errata says that the errata "are not to be considered normative until approved ..." These are essentially proposals which if accepted will then be published as the 5th edition.
(In reply to comment #4) > The status section of the 4e errata says that the errata "are not to be > considered normative until approved ..." > > These are essentially proposals which if accepted will then be published as the > 5th edition. > oops yes you are right, but the main point of comment #3 stands, once this change _is_ approved, it will apply normatively by errata to earlier editions, so just referencing a dated version would not be sufficient, you'd have to explicitly (n words) refer to the edition as published, unmodified by errata. David
and it's approved now: http://www.w3.org/TR/2008/REC-xml-20081126/ Jonathan
Looks like we can simply eliminate that choice now. Jonathan
At the 2009-02-05 telcon the XSL WG decided (subject to confirmation by the XQuery WG) to adopt for XPath the analogous resolution that the XSLT XQuery joint telcon decided to adopt for XQuery, namely: XPath 2.0: The version of XML 1.0 must be no earlier than the third edition; the exact edition is implementation-defined, but we recommend that implementations use the latest XML 1.0 version. XPath 2.1: Not refer to any specific edition (implying the latest).