I suspect its aX+b where X is the source size and b is a constant. That is, it's linear but not proportional.
 
Michael Kay
http://www.saxonica.com/


From: saxon-help-bounces@lists.sourceforge.net [mailto:saxon-help-bounces@lists.sourceforge.net] On Behalf Of Edelson, Justin
Sent: 23 October 2006 17:20
To: Mailing list for SAXON XSLT queries
Subject: Re: [saxon] Size of Templates objects

I would guess (and to be clear, I've done absolutely nothing to validate this) that it'd be better than linear because of overhead common to all stylesheets regardless of size. But I didn't write the thing :)
 
Thanks,
Justin


From: saxon-help-bounces@lists.sourceforge.net [mailto:saxon-help-bounces@lists.sourceforge.net] On Behalf Of Michael Kay
Sent: Monday, October 23, 2006 12:13 PM
To: 'Mailing list for SAXON XSLT queries'
Subject: Re: [saxon] Size of Templates objects

To be honest, the size of the Templates object hasn't had much attention because it is rarely significant. It's the run-time memory usage that people worry about.
 
But I would expect it to be essentially linear with source code size. When you say you think this isn't the case, is your impression that it is better than linear or worse than linear?
 
Michael Kay
http://www.saxonica.com/


From: saxon-help-bounces@lists.sourceforge.net [mailto:saxon-help-bounces@lists.sourceforge.net] On Behalf Of Edelson, Justin
Sent: 23 October 2006 17:02
To: Mailing list for SAXON XSLT queries
Subject: [saxon] Size of Templates objects

Does anyone know how the memory usage of a JAXP/Saxon Templates objects scale in relation to the size/complexity of the underlying stylesheet?

In other words, does a 2000-line stylesheet take up twice the memory as a 1000-line stylesheet? My gut is no, but my gut's been wrong before.

I'm seeing some memory consumption issues which may or may not be XSLT-related and before I created some test cases to profile, I thought I'd ask this list to see if anyone has done this type of analysis before.

Thanks,
Justin