RE: Could XQuery replace the W3C DOM?

Precisely my point. That sort of performance is going to be optimal for
known repetitious queries (which in the real world we know to make up a huge
percentage of queries in applications. PS. How do I learn more about this,
sounds cool!

-----Original Message-----
From: Per Bothner [mailto:per@bothner.com]
Sent: Thursday, August 29, 2002 1:32 PM
To: Adam Bosworth
Cc: 'noah_mendelsohn@us.ibm.com'; www-ql@w3.org
Subject: Re: Could XQuery replace the W3C DOM?


Adam Bosworth wrote:
>>From a performance point of view, one can imagine a factory object that
> given xml queries, generates optimal code for executing them.

Well, that dort of what Qexo does:  It compiles Java bytecodes from
an XQuery "program", and can do so on-the-fly.

> -----Original Message-----
> From: noah_mendelsohn@us.ibm.com [mailto:noah_mendelsohn@us.ibm.com]
 >
> DOM is already somewhat slow/heavyweight for certain purposes.

Qexo uses it's own lightweight DOM, where each document (or document
fragment) is encoded in a flat array of chars.  This is a binary,
not textual, encoding, using relative offsets to point from a
begin element "chunk" to the corresponding end, and vice versa.
A Node is represented an an int offset into this array.  An entire
document is one TreeList object plus this char array plus an array
of Object nodes.  (For example an element's QName is an offset
in this Object array.)  In contrast, standard DOM uses one object
per node, and each object has lots of fields.
-- 
	--Per Bothner
per@bothner.com   http://www.bothner.com/per/

Received on Friday, 30 August 2002 08:48:37 UTC