W3C home > Mailing lists > Public > www-qa-wg@w3.org > June 2004

Re: Response: Two more comments on "SpecLite"

From: Jeremy Carroll <jjc@hplb.hpl.hp.com>
Date: Fri, 18 Jun 2004 16:10:12 +0100
Message-ID: <40D305D4.1010905@hplb.hpl.hp.com>
To: Lynne Rosenthal <lynne.rosenthal@nist.gov>
Cc: www-qa-wg@w3.org

Correct numbers:

2 required

9 not recommended

34 optional!

(note this includes rdf:XMLLiteral as an optional, so the total should 
be one more than the number of XML Schema builtin simple datatypes)

(I have opted for "seventeen billion" rather than
"17179869184" as the value of 2^34)


>> [[
>> One WG decided to use the built-in datatypes from XML Schema in their
>> specification. However, on inspection it become clear that a few of
>> these were inappropriate, and that some implementors thought the cost of
>> implementing all of the rest prohibitive. 

Thus the WG identified: nine datatypes as not recommended,
and two as required,
leaving the
other thirty-three as optional. They failed to identify subsets of the
>> optional datatypes, leaving implementors free to choose any of the 
seventeen billion
possible subsets. The WG never articulated the extent of this freedom,
>> nor wondered whether interoperability would be enhanced by restricting
>> it in a sensible manner. Until they reached PR they also failed to
>> consider the interaction between this and the profiles that their
>> specification included. One of these profiles extended another
>> specification that required the implementation of one of the optional
>> datatypes. One AC reviewer drew critical attention to the potential
>> conflicts in the datatypes mismatch at the extension point, resulting in
>> heated discussion in the WG and a last minute patch to the specification
>> in the last few weeks before it went to Recommendation.
>> ]]
Received on Friday, 18 June 2004 11:10:29 GMT

This archive was generated by hypermail 2.2.0 + w3c-0.30 : Thursday, 9 June 2005 12:13:16 GMT