RE: [xml-dev] Concerned about the increasing reliance on XPath

["David Lee" <dlee@calldei.com>]

==> Liam
"They're in memory except when they aren't"?  The Working Groups
designing XPath 2, XSLT 2 and XQuery 1 were very aware that the data
model instances over which these languages operate might not be entirely
in main memory...
======

Yes exactly.

There are no absolute statements.   (even this one). 
I was presuming that Roger was asking about non-XML DB or more standard
XPath implementations.
So yes "They're in memory except when they aren't"
The *vast majority* of *full* XPath implementations that I am aware of
require all data to be in memory.
And when they are not many XPath statements cannot be implemented in
streaming mode without some very advanced coding more typical in XML DB's ..
in which case indexes take the place of memory.
I won't argue (today) about what these are, its  well-known and published
fact.
( Of course there are *subsets of XPath* which can be streamed - but the
whole-shebang really cant).

Even if the working group was aware that the data may not be in memory
doesn't mean that implementations can easily support all of XPath (even 1.0)
without the data being in memory
(or indexed, hence it was pre-processed or the data was read in its entirety
with all possible xpaths to be executed know ahead of time. ).

Of course "streaming" is not exactly the same as "not in memory" but that
would be a LONG topic ...



==> Liam
We do agree here ;-)

I think we agree on all points.


----------------------------------------
David A. Lee
dlee@calldei.com
http://www.xmlsh.org