W3C home > Mailing lists > Public > www-tag@w3.org > May 2009

Re: Comment on XSD 1.1

From: Pete Cordell <petexmldev@codalogic.com>
Date: Fri, 22 May 2009 12:02:40 +0100
Message-ID: <27DAAB4293714B1F9D925E08BEDD6B82@Codalogic>
To: <www-tag@w3.org>
Cc: <www-xml-schema-comments@w3.org>
Original Message From: "Michael Kay"
>> For the interest of the TAG group:
>>     http://www.w3.org/TR/2004/REC-xmlschema-1-20041028/
>> 137.62 KB (140,924 bytes)
>>     http://www.w3.org/TR/2009/CR-xmlschema11-1-20090430/
>> 285.19 KB (292,037 bytes)
>> Why is XSD 1.1 Part 1 twice the number of characters compared
>> to XSD 1.0?  Both use UTF-8. (It presumably cannot be mostly
>> the change list of Annex G.)  Richer markup?
> The number of characters in non-whitespace text nodes in the XHTML has
> increased from 442541 to 827066, so yes, this is a genuine increase in the
> length of the spec. However, length does not equal complexity.

>From my perspective one of the main benefits I'm hoping for from XSD 1.1 is 
a clearer spec.  I'll confess I haven't read a draft lately, but I'm not 
worried about it getting longer as long as it gets clearer.  And while the 
Relax-NG spec is shorter than XSD, Relax-NG's formula based description of 
things would surely increase the length of the spec if they were presented 
in narrative form.  I do find the Relax-NG spec difficult to read also. 
There's always seems to be a nagging feeling of have I really understood 
what that means.

Personally I would have preferred XSD1.1 to be more of a maintenance 
version, fixing what seemed clearly wrong, rather than adding new 
functionality.  To that end I'd have gone for:

- Relax the UPA constraints for wildcards.

- Relax cardinality constraints of xs:all members.

- Possibly introduce something like xs:alternate but limit the test to 
"AttrName '=' Literal"

That would cover 90% of the problems I see people having with XSD today.  An 
XSD 1.2 (or something new) could then add additional features such as 
xs:assert and so on.

On whether we should approve XSD 1.1 or do something new, I don't see it as 
either/or.  I think there is space for both approaches.  XSD doesn't give 
much bang for the buck, but the bucks have mostly been paid now, and doing 
something else would largely involve paying more bucks than saving bucks.


Pete Cordell
Codalogic Ltd
Interface XML to C++ the easy way using XML C++
data binding to convert XSD schemas to C++ classes.
Visit http://www.xmldatabinding.net for more info.
Received on Friday, 22 May 2009 11:03:49 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 22:56:29 UTC