Hey,
On 15 Apr 2005, at 5:38, Gregor wrote:
Hi,
thank you all for these interesting thoughts!
Jeff had the idea of using the grammar preparsing mechanism to validate
schemas. I like the idea from the point of view that I want to benchmark
Xerces validation performance in a B2B scenario.
Therefore precaching is imho required anyway for increased performance in
message validation.
Caching will improve performance if repeated validation occurs. From the benchmarking perspective it can make sense, as long as you don't want the time taken to parse the schema included in the benchmark. As Jeff points out, preparsing them will check them to xerces standards.
The benchmark however relies on the XML Schema test collection available
here (http://www.w3.org/2001/05/xmlschema-test-collection.html). In the
past, xerces (2.0.0_beta4) was only tested in test cases that involved a
schema and an instance.
What kind of a benchmark do you want to create? If it is instance validation then you can cache all those schemas and and time how long it takes for xerces to validate the instances against them. If you want to benchmark how quickly xerces parses and extracts info for schema validation you can time the caching process itself.
Please share with me your thoughts on:
1. ) Did anybody try to run the W3C schema-only test cases on Xerces (using
the precaching framework)?
Don't know. IBM have extensive schema tests that get run on xerces prior to release (or more). These may well be part of that.
2. ) Is the precaching framework suitable for running the above benchmark?
See above comments.
3. ) Will Schema validation be added to Xerces (the last question is interesting for the "outlook"-part of my paper since I want to compare Xerces with an OpenSource Tool developed by my company).
I know of no plans to fully implement the constraints of the schema specs (that does not mean it is not happening :) ). If you think it would be useful functionality, over and above other things, then speak up. If enough people say then it might happen.
Thank you Gareth, Jeff and Bob for your previous answers.
Kind Regards,
Gregor
@Gareth: This is how I came up with the idea of using the Schema for Schema
with Xerces (it was silly, but I gave it a try :-)
From XSV homepage (http://www.w3.org/2001/03/webdata/xsv)
Its a completely reasonable approach, but as much with the schema specs, it is not always totally clear to mere mortals like us :)
Regards,
Gareth
-- Gareth Reakes, Managing Director Parthenon Computing +44-1865-811184 http://blog.parthcomp.com/xerces
--------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
