- From: Don Chamberlin <chamberl@almaden.ibm.com>
- Date: Mon, 23 Jun 2003 11:33:05 -0600
- To: xquery@attbi.com, public-qt-comments@w3.org
- Message-ID: <OF61EF2998.C33E8725-ON88256D4E.005EB4A2-88256D4E.00600E19@us.ibm.com>
Michael,
Thanks for your suggestions. I'll fix the errors in the next edition of
the document.
Specifically:
(1) I will add the missing curly braces to the example document node
constructor in Section 3.7.2.3. I believe that the example will then be
correct. The path expression doc("bib.xml")//book/author returns a
sequence of zero or more author elements, not a document node.
(2) In the first example of Section 3.7.2.1, the use of the function
node-name($e) to compute the node name of the constructed element is an
example of a reasonable and valid query that is nevertheless considered to
be a "type error" by our conservative rules for static type checking.
There are a great many such examples. That is why type checking is
optional in XQuery. A system that performs dynamic type checking will
execute this query without any problem. A system that performs only static
type checking will be unable to execute this useful query.
(3) In the second example of Section 3.7.2.1, I will insert a "data"
function call into the name expression to fix the bug you have noted. The
name expression will then look like this:
{data($dict/entry[word=name($e)]/variant[lang="Italian"])}
Regards,
--Don
Received on Monday, 23 June 2003 13:33:15 UTC