- From: Michael Kay <mike@saxonica.com>
- Date: Sun, 13 May 2007 22:13:24 +0100
- To: "'Le, Yongnian'" <yongnian.le@intel.com>, <xmlschema-dev@w3.org>
- Cc: "'Jones, Kevin'" <kevin.jones@intel.com>, "'Yu, Zhiqiang'" <zhiqiang.yu@intel.com>
- Message-ID: <021c01c795a3$8b8ba0a0$6401a8c0@turtle>
You misunderstood me. I didn't say that there were no real applications using identity constraints. I said that I wasn't aware of any in the public domain that could usefully be used for benchmarking. As far as I can see you are conjecturing performance difficulties here with no real evidence. I don't think there is any reason to believe that the cost of validating identity constraints is likely to be a significant obstacle to real applications. Michael Kay http://www.saxonica.com/ _____ From: Le, Yongnian [mailto:yongnian.le@intel.com] Sent: 13 May 2007 17:17 To: Michael Kay; xmlschema-dev@w3.org Cc: Jones, Kevin; Yu, Zhiqiang Subject: RE: is there any formal commercial identity constraint cases That’s a pity that there is no real application having identity constraints, maybe because that feature is quite advanced. Anyway, thanks for your info. The differences between identity constraint and other regular expressions are mostly listed below, which makes me concern the runtime overhead of identity constraint support in the real application usage. 1. Identity constraint processing creates dependence between validations in different SAX events, which make it hard for further advanced optimization; while regular expression in schema validation can be processed in a single SAX event validation. 2. Identity constraint requires typed based comparison, which might involve a lot of table/value-object/memory management, type conversion and string operations during validation; while regular expression processing is relatively simple: accepted or not accepted. Best Regards Yongnian _____ From: Michael Kay [mailto:mike@saxonica.com] Sent: 2007年5月11日 15:49 To: Le, Yongnian; xmlschema-dev@w3.org Cc: Jones, Kevin; Yu, Zhiqiang Subject: RE: is there any formal commercial identity constraint cases Thanks for the explanation. I'm not aware of any public domain material that can be used for performance benchmarking of schema processors, whether it's concerned with identity constraints or anything else. Is there any particular reason why you think identity constraints should be a problem, as distinct from, say, validation of strings against complex regular expressions? Michael Kay http://www.saxonica.com/ _____ From: Le, Yongnian [mailto:yongnian.le@intel.com] Sent: 11 May 2007 02:43 To: Michael Kay; xmlschema-dev@w3.org Cc: Jones, Kevin; Yu, Zhiqiang Subject: RE: is there any formal commercial identity constraint cases I mean that whether there is any real customer case which contains identity constraint for us to use that as benchmark to measure performance of identity constraint support. Currently I only found there are hundreds of identity constraint unit cases in w3c XML Test Suite, but I am afraid that all of them are not representative enough in the real world used by customer. Best Regards Yongnian _____ From: Michael Kay [mailto:mike@saxonica.com] Sent: 2007年5月10日 17:42 To: Le, Yongnian; xmlschema-dev@w3.org Cc: Jones, Kevin; Yu, Zhiqiang Subject: RE: is there any formal commercial identity constraint cases I'm not sure what you mean by "formal", "commercial", or "benchmarking". Michael Kay http://www.saxonica.com/ _____ From: xmlschema-dev-request@w3.org [mailto:xmlschema-dev-request@w3.org] On Behalf Of Le, Yongnian Sent: 10 May 2007 09:52 To: xmlschema-dev@w3.org Cc: Jones, Kevin; Yu, Zhiqiang Subject: is there any formal commercial identity constraint cases Hi everyone, Do you know whether there is any formal commercial identity constraint case available for benchmarking? Thanks a lot! "I hear and I forget, I see and I remember, I do and I understand" -- quoted by Yongnian ICSC Compiler Le, Yongnian 8752-6253
Received on Sunday, 13 May 2007 21:13:44 UTC