Re: [xml-dev] XML Schema Checklist

[Timothy Cook <timothywayne.cook@gmail.com>]

Hi Ken & All,

On Sat, Oct 20, 2012 at 2:29 PM, G. Ken Holman
<gkholman@cranesoftwrights.com> wrote:
> At 2012-10-20 13:17 -0300, Timothy Cook wrote:
>> instance documents on them. There should be some way for a parser to
>> know which schema version it is parseing. Other than waiting for it to
>> "bump into" a 1.1 capability.
>
>
> Actually, I think that's about it.  Certainly it is true for XSLT and it is
> designed into XSLT that unrecognized constructs are innocuous in a
> stylesheet unless they are encountered.
>

I must say that I am more than a little surprised.  But ...

So, if I can bother with one last point on this issue.  Since I
apparently misunderstood   1.3.4
(http://www.w3.org/TR/xmlschema11-1/#langids)

"Sometimes other specifications or Application Programming Interfaces
(APIs) need to refer to the XML Schema Definition Language in general,
sometimes they need to refer to a specific version of the language,
possibly even to a version defined in a superseded draft. To make such
references easy and enable consistent identifiers to be used, we
provide the following URIs to identify these concepts."

It seems pretty clearly written to me.  However, if I did understand
it.  Then it is indirect conflict with your (Ken's) earlier statement:

<<<
 And, normatively, section 1.3.1.1 unambiguously states that the
version-agnostic URI shall be used:

  http://www.w3.org/TR/2012/REC-xmlschema11-1-20120405/#xsd-namespace

So I do not see any interpretation of the specification that would
allow what you ask.
>>>

I do not see the SHALL part you mention.  In fact I get just the
opposite impression.
This cause me to believe that the version-agnostic URI was used in the
specification just to be consistent, not normative.

>> I know that this has been thought of, so what am I still missing?
>
>
> I'm scanning the W3C Schema 1.0 document and cannot quickly put my finger on
> the chapter and verse that addresses a schema expression containing non-1.0
> constructs.  Perhaps someone else on the list knows where this is found.
>

Perhaps.  However, the non-normative section
http://www.w3.org/TR/2012/REC-xmlschema11-1-20120405/#nonnormative-language-ids
seems to me to support my understanding of 1.3.4

If my understanding of 1.3.4 is incorrect.  Then can someone tell me
what it does say?

Thanks,
Tim

P.S.  Apologies if this line of questions seems naive to those with
more experience.  But I think it is an important one to understand.  I
am curious how tool builders approach this issue. SAXON?  XERCES?


============================================
Timothy Cook, MSc           +55 21 94711995
MLHIM http://www.mlhim.org
LinkedIn Profile:http://www.linkedin.com/in/timothywaynecook
Academic.Edu Profile: http://uff.academia.edu/TimothyCook
Google Scholar: http://goo.gl/MMZ1o