Are your referring to the Java processor or the C++ processor?
Dave
|---------+--------------------------->
| | "Vitaly |
| | Korovinsky" |
| | <vitalyk@sympati|
| | co.ca> |
| | |
| | 04/15/2002 01:44|
| | PM |
| | Please respond |
| | to xalan-dev |
| | |
|---------+--------------------------->
>---------------------------------------------------------------------------------------------------------------------------|
|
|
| To: <[EMAIL PROTECTED]>
|
| cc: (bcc: David N Bertoni/Cambridge/IBM)
|
| Subject: Using Xalan in a database environment
|
>---------------------------------------------------------------------------------------------------------------------------|
I am trying to use Xalan for the data that is stored not in an XML file,
but
in the database. There is too much of data to be loaded as a transient DOM
tree into physical memory. We are trying to find a way of using Xalan when
only a few nodes are loaded into the memory, but others could be fetched on
demand.
It seems to me that Xalan's DOM model is built on the assumption that the
whole DOM tree is loaded into the memory, therefore interface can return a
pointer to a child node or reference to a string. However if the DOM tree
is
so large that it can't be held in memory a different interface is required.
The one that uses proxy objects instead of pointers. A proxy object would
serve as a smart pointer and load object's details from the database on
demand.
Is anyone out there trying to work with very large DOM trees? Does anyone
knows whether Xalan will address the issue of large DOM trees in the near
or
remote future? Or maybe someone knows that Xalan should solve other
improtant issues first and dealing with large DOM trees will not be
addressed any time soon? ;-)
Thanks a lot,
Vitaly
