17603 2012-06-26 13:14:58 +0000 [XQ31ReqUC] Change semantics of identity comparison and attempt to make XQuery pure functional 2014-05-20 17:03:02 +0000 1 1 1 Unclassified XPath / XQuery / XSLT Requirements for Future Versions Working drafts All All NEW P2 normal --- 1 g jim.melton jim.melton public-qt-comments oldest_to_newest 69407 0 g 2012-06-26 13:14:58 +0000 It would be nice to consider Tim's suggestion about changing the definition of node identity comparison to no longer expose the physical identity of nodes. https://lists.w3.org/Archives/Member/w3c-xsl-query/2012Jun/0058.html I see two advantages of this approach: (i) it addresses the optimization issue related to the non-determinism of functions returning newly created nodes (ii) it would remove one reason why XQuery is not pure functional (if I am correct, this would leave us with error handling only). In particular, variable substitution would become possible even in the presence of new nodes. 106487 1 jonathan.robie 2014-05-20 17:03:02 +0000 Assigning to future requirements per Working Group decision (https://lists.w3.org/Archives/Member/w3c-xsl-query/2012Oct/0087.html).