[XML-DEV Mailing List Archive Home] [By Thread] [By Date] [Recent Entries] [Reply To This Message]

RE: XML-enabled databases, XQuery APIs

• To: 'Ken North' <kennorth@s...>, xml-dev@l...
• Subject: RE: XML-enabled databases, XQuery APIs
• From: "Bullard, Claude L (Len)" <len.bullard@i...>
• Date: Wed, 20 Apr 2005 11:23:48 -0500

Thanks Ken, for that referenced paper.  Very good reading.  And

"Our experiments have shown that there is some merit to suggestions 
that XML parsing and application performance can be improved upon 
through binary encodings.  However, more significant improvements 
can be otatined by adopting optimized stream based XML parsers if 
they are not already being used.  Also, because C and C++ parsers 
outperform their Java counterparts, Java applications could make 
substantial performance gains through native libraries for handling 
the XML streams.

We have also found that some optimizations such as tokenization 
do not lead to substantial performance improvements even if the 
application can exploit tokens for reduced string comparison cost. 
Tokenization may be more useful in applications requiring XML 
compression though further experiments are required to justify 
it even in this context since generic compression schemes may be 
just as effective."

Hmm.  So, a trivial binary standard is the sweet spot for 
standardization; compression is a separate issue needing further 
study, and schema-based binarization tightly couples but is 
possibly the best approach in the cases where high performance 
is traded against tight coupling (effectively, a new media type) 
and the view source principle (in cases where the media type 
need not be mastered by viewing the work of others as published), 
and more speed can be had by native libraries written in faster 
languages.

The claim that view source is the reason for the growth of the 
web is overhyped.  Applications such as Flash became ubiquitous 
without it and making it view sourceable now doesn't change the 
historical facts.  For some applications, the capabilities and 
any associated religious convictions about them will be just as 
powerful.  For evidence, I offer the ubiquity of Flash and the 
refusal of VRML to die. Rationality is a weak predictor of 
human behavior or choice because some people don't know about 
the rationale or don't care.  They want what they want.

- Logic is a plug-in processor in the mammal brain.  It 
  is dominated by emotional processing.
- Some people don't care about the strategies.
- Some people don't share the values.

len

• Follow-Ups:
  • Re: XML-enabled databases, XQuery APIs
    • From: Michael Champion <michaelc.champion@g...>
  • A Stakeholder's Response: XQuery APIs for Middle Tier and Clients
    • From: Jonathan Robie <jonathan.robie@d...>

• Prev by Date: Re: adding markup to tables that have none?
• Next by Date: newbie help -- schema development
• Previous by thread: Re: XML-enabled databases, XQuery APIs
• Next by thread: A Stakeholder's Response: XQuery APIs for Middle Tier and Clients
• Index(es):
  • Date
  • Thread