1790 2005-07-20 20:47:27 +0000 [FS] technical: normalization rules "not required" 2005-09-27 08:42:48 +0000 1 1 1 Unclassified XPath / XQuery / XSLT Formal Semantics 1.0 Last Call drafts All All CLOSED FIXED http://lists.w3.org/Archives/Public/public-qt-comments/2002May/0056.html P2 normal --- 1 jmdyck simeon public-qt-comments oldest_to_newest 5030 0 jmdyck 2005-07-20 20:47:27 +0000 Various sections 4.1.1 Literals 4.1.2 Variable References 5.4 Default Collation Declaration 5.6 Construction Declaration 5.10 Schema Import 5.12 Namespace Declaration 5.13 Default Namespace Declaration say something like: All X are Core expressions, therefore no normalization rules are required for X. I disagree that they are not required. The normalization of 'higher-level' constructs (e.g., an Expr, or a PrologDeclList) depends on (is defined in terms of) the normalization of its sub-constructs. If you don't provide a mapping rule for the sub-construct, then that normalization isn't defined, and the higher-level normalization fails. No big deal, just replace the above wording with: [[ X ]]_whatever == X Note that some sections *do* supply identity normalization rules, e.g.: 4.2.1.1 / Norm (under "all other forward/reverse axes") 4.7.1 / Norm / rule (6|7) 5.15 / Notation / rule 2 (This is a leftover from 3 years ago. See attached URL.) 5033 1 simeon 2005-07-20 21:35:18 +0000 The working group aggrees with you. We will add the identity normalization rules for the cases you mention. - Jerome