[
Lists Home |
Date Index |
Thread Index
]
Jonathan Robie wrote:
> If an operand of an arithmetic operator is not typed, XQuery attempts to
> cast it to a double. This cast may fail. This is exactly what XPath does,
> and it is also - as Joe points out - analogous to what TCL does.
>
> Of course, if datatypes are present, they are observed, unless the writer
> of the query explicitly casts data to something untyped.
Hi Jonathan.
I am afraid that you and I are once again talking past each other. I am aware
that this is what XPath does. I believe that this already represents too much
'direct access' through an implicit datamodel. As Dare has pointed out
(http://lists.xml.org/archives/xml-dev/200301/msg00426.html)--with, I
believe, approval on his part where there is dismay on mine--"XPath is only
meant to run against virtual XML views and not UnicodeWithAngleBrackets". I
believe that *every* use of XML (demonstrating, inter alia, that it *is* XML)
requires a parse of UnicodeWithAngleBrackets and then upon the output of that
parse the instantiation of whatever data structure is required for execution
of the particular function at hand. To me it is axiomatic that access to XML
can never be any more direct than that. Others differ in their opinions.
Best regards,
Walter
|
