No you will actually loose performance if you include more data.
A DTD is inarguably more data so you will loose performance.
If you do not need validation (and you seem not to) then do not use it.

We did a simple benchmark on validation vs no-validation when we started
using Xerces. And the performance hit is roughly 10% with validation on. But
that was in the old DOM. Not quite sure how the figures is in the IDOM
environment.
But our design decision was to use validation anyway do be sure that each
document arrives intact through compession, crypto and network.
But that decision is hanging by a thin thread. If I need the performance
then the DTD gets axed in a heartbeat (we also have a controlled environment
and I have yet to see a corrupted package)

Regards

Erik Rydgren
Mandarinen systems AB
Sweden


-----Original Message-----
From: Bagepalli, Kiran [mailto:[EMAIL PROTECTED]]
Sent: den 16 januari 2003 01:13
To: '[EMAIL PROTECTED]'
Subject: loadGrammar


I would like to know the performance benefits of loadGrammar.

Actually, I have a lot of small XML files all with the same structure and
has no DTD or schema reference.
It takes a while to load the files. I am creating multiple DOMParsers and
scanning each document.

Would the same be faster, if I have a DTD reference in the XML files and
preloaded the grammar and then scanned the same files.

Today I am having no DTD reference since the XML is internal and the
structure is fixed.

Thanks
Kiran

---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]


---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]

Reply via email to