cocoon-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Marc Portier <>
Subject SAX done better?
Date Thu, 13 Nov 2003 21:52:07 GMT
Hi all,

just been reading up on the new StAX API 

can't help being (a bit) enthousiastic, looks to me like an important 
step into the direction of efficient parsing of XML
(seems to be more valuable then just a symbiosis of the existing pull 
parsing approaches currently out there)

the spec is short and provides a genuine brain-feeding read: in fact one 
gets to notice how SAX left some room for efficiency improvements: e.g. 
the characters event in SAX cannot be trusted not to reuse the same 
buffer variable which means you have to do a mem-copy on the event even 
if only later in the event-stream you can decide that you could discard 
it anyway...

still have to do some personal tests to really dig what they mean with 
their proclaimed pipelining support, but already I have the idea that 
this is one to keep an eye on in our line of working (e.g. FO 
implementations could benefit IMHO)

also their (to be done in future release :-() 'virtual data source' is 
read by me as some natural fit on our pseudo-protocols meeting 
java-beans passed/stored in the request, flow or session context

well, if nothing else, the trendspotter in me kind of noticed this:

# Support for non-namespace aware documents becomes optional!
==> You are running out of excuses to not be doing namespaces!
# XML's verbosity is calling upon some efficiency focus
==> XML starts being used in domains with large datasets or documents
# SAX-meme is infecting XML-designs
==> Sensible improvements (more avoided mem-copies)
==> Another API endorsing the XML-Pipeline-Processing paradigm

Marc Portier                  
Outerthought - Open Source, Java & XML Competence Support Center
Read my weblog at                        

View raw message