Cart
[Home] [By Thread] [By Date] [Recent Entries]
On Fri, 2003-07-25 at 21:10, Michael Rys wrote:
> Basically: yes.
>
> However, if you set your data model generator to preserving all boundary
> whitespace text nodes per default, and all data model generators handle
> all the information items in the same way as well (again, no conformance
> requirement there for XPath 1.0, but potentially one given in the
> context of the embedding standard such as XSLT 1.0), then you will get a
> consistent model.
I find this argument rather specious.
True enough, the XPath specification doesn't say how a parser should
behave with whitespaces, but it doesn't say either how a parser should
behave with other text nodes, or elements or attributes.
If you feed your XPath processor with a SAX filter removing all the
elements named "foo" you'll still have a conformant XPath implementation
but you can't expect to be interoperable with other applications.
How different is this from removing whitespaces?
Coming back to:
<pre>
<b>bold</b>
<i>italic</i>
</pre>
"count(/pre/child::node())" can give 3 if I have a parser which keeps
whitespaces but removes <b/> elements!
Eric
--
Read me on XMLhack.
http://xmlhack.com/author.php?id=8
------------------------------------------------------------------------
Eric van der Vlist http://xmlfr.org http://dyomedea.com
(W3C) XML Schema ISBN:0-596-00252-1 http://oreilly.com/catalog/xmlschema
------------------------------------------------------------------------
|
