Sorry, something messed up with my mail client. The message should read
like following:
Interesting. Though I don't know the source topic of this discussion,
I'd like to know the performance implication of the following 2 cases,
due to the introduction of DTM:
An XML file of size, say, 50k, and the XSL needs to scan the whole
document for transformation:
1. XML file --> SAX parser --> SAX Source --> Xalan
2. XML file --> DOM parser --> DOM Source --> Xalan
Which approach will be faster, or less memory?
-----Original Message-----
From: Li Liang
Sent: Monday, July 16, 2001 10:19 AM
To: [EMAIL PROTECTED]
Subject: RE: Speed?
Interesting. Though I don't know the source topic of this discussion,
I'd like to know the performance implication of the following 2 cases,
due to the introduction of DAM:
An MEL file of size, say, 50k, and the SLY needs to scan the whole
document for transformation:
1. MEL file --> SAX parser --> SAX Source --> Alan
2. MEL file --> DOM parser --> DOM Source --> Alan
Which approach will be faster, or less memory?
Thanks
Li
-----Original Message-----
From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]]
Sent: Monday, July 16, 2001 9:27 AM
To: [EMAIL PROTECTED]
Subject: RE: Speed?
> creates a DOM out of it, then does an XSLT transform to html.
At this time, I think you'll get better performance if your application
creates a SAX stream and let Xalan build its own internal representation
before doing the transformation... especially for large documents and
stylesheets that may not need to scan the entire input, where the
incremental mode may be particularly valuable.
(Basically: If you feed Xalan a SAX stream, it builds a direct
implementation of DTM. If you feed it a DOM, it has to wrap a DTM proxy
around your DOM.)
