Cart
[Home] [By Thread] [By Date] [Recent Entries]
> NOT reading/parsing and rereading/reparsing relatively static > XML documents gives me a huge performance boost. I was under impression that it should usually give about 10-15% boost. Is that 'huge' ? Well, maybe I don't understand something... > In fact that is the major reason why I moved from SAX > based XSLT rendering to DOM based rendering for the vast majority of our > stuff. XSLT rendering is not 'SAX based'. It is 'OM-based'. XSLT keeps the entire document in the memory. Well, there are some XSLT engines that try to do something smarter, ( like SAXON's 'preview' ) but I don't think that using 'preview' gives any 'speed boost' . In the cases it *does*, that means that the document is anyway too huge to be processed by DOM at all. In the case when 10-15% of speedup is what you're looking for, something like Chunks [1] or other Brutal XML bindings [2] should outperform DOM significantly. My prediction is that the era of low-level lexer ( called SAX ) and low-level model ( called DOM) is over and there will be soon more high-level bindings on top of these low-level APIs (or not on top of them). I think that asking developers to write all the code in terms of SAX or DOM APIs is like asking them to write programms in assembly language. Rgds.Paul. [1] http://www.pault.com/pault/pxml/nxml.html [2] Here are two other Brutal XML Bindings http://web.co.nz/~grantm/cpan/index.html http://www.zope.org/Members/haqa/XMLKit I appreciate more URLs for brital XML bindings, seems not easy to find them ...
|
