[Bug 1380] New: [XQuery] some editorial comments on A.1.1 grammar-note: xml-version


           Summary: [XQuery] some editorial comments on A.1.1 grammar-note:
           Product: XPath / XQuery / XSLT
           Version: Last Call drafts
          Platform: All
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: XQuery
        AssignedTo: chamberl@almaden.ibm.com
        ReportedBy: jmdyck@ibiblio.org
         QAContact: public-qt-comments@w3.org

A.1.1 grammar-note: xml-version

"For convenience, XML 1.0 references are always used."
    It isn't clear what the scope/meaning of that sentence is.  Maybe change to
    "The EBNF only has references to the 1.0 versions."

    It would be nice to have links to the 1.1 versions too.
    Maybe the EBNF section should have something like this:
        It is implementation-defined whether to use these:
            [143] PITarget :: = [http:...1.0...]
            [148] CharRef  :: = [http:...1.0...]
            [156] QName    :: = [http:...1.0...]
            [157] NCName   :: = [http:...1.0...]
            [158] S        :: = [http:...1.0...]
            [159] Char     :: = [http:...1.0...]
        or these:
            [143] PITarget :: = [http:...1.1...]
            [148] CharRef  :: = [http:...1.1...]
            [156] QName    :: = [http:...1.1...]
            [157] NCName   :: = [http:...1.1...]
            [158] S        :: = [http:...1.1...]
            [159] Char     :: = [http:...1.1...]

    The external definitions of CharRef both have a well-formedness constraint.
    Does this apply to occurrences of CharRef in XQuery?  And if so, what error
    occurs if a CharRef satisfies the EBNF but fails the WFC?

"In some cases, the XML 1.0 and XML 1.1 definitions may be exactly the same."
    Why be coy?  Presumably, the spec can say which are the same.  As far as I
    can see, for PITarget, CharRef, and S, the two definitions are the same, and
    for QName, they're equivalent.  (Leaves NCName and Char.)

    If the 1.0 and 1.1 definitions are the same, why not simply put the
    definition in the XQuery spec, instead of referring to it in two external
    specs? Well, I can see why you might not want to do it for PITarget, because
    you'd then need to have a production for 'Name', which might be confusing.
    But for CharRef, S, and QName, I don't see a downside. In fact, for CharRef,
    there's the added bonus that you wouldn't have to talk about how an external
    XML WFC relates (or doesn't) to XQuery, you could just express it as a
    grammar-note (or not).

"just as it is not permitted in a textual XML document."
    Delete "textual" ? (Is there any other kind of XML document?)

Received on Wednesday, 11 May 2005 07:33:19 UTC