RE: Xml Schema profile

Indeed you are correct.  If all folks ever want is simple document validation, then their profile may not restrict very
much.  However, I suspect there would still need to be a profile.  Take redefine.  Support for this troublesome feature
isn't ubiquitous even in mature, validating parsers.  

But I get your point and I agree.  It is hard to get consensus on "unnecessary" features, so perhaps my postings are the
consensus of our organization only.

Thanks for your comments.





 

-----Original Message-----
From: Michael Kay [mailto:mike@saxonica.com] 
Sent: Wednesday, July 19, 2006 12:58 PM
To: 'Paul Kiel'; xmlschema-dev@w3c.org
Subject: RE: Xml Schema profile

I think previous attempts to identify what could be left out achieved little
consensus: it seems it's rather like Microsoft Word: everyone thinks there are too many features, but no-one agrees
which features are unnecessary.

You seem to be approaching this from the perspective of code-generation tools. These notoriously tend to leave out the
features that don't have ready-made equivalents in conventional programming languages. But for people using XML Schema
for its original purpose, namely document validation, these features are extremely important.

Michael Kay
http://www.saxonica.com/

> -----Original Message-----
> From: xmlschema-dev-request@w3.org
> [mailto:xmlschema-dev-request@w3.org] On Behalf Of Paul Kiel
> Sent: 19 July 2006 17:45
> To: xmlschema-dev@w3c.org
> Subject: Xml Schema profile
> 
> 
> Hi Folks,
> 
> I know that in some circles the idea of creating a profile for Xml 
> Schema meets with politically strong feelings.  There are those that 
> argue that we should shame tools makers into supporting 100% of the 
> Xml Schema spec.  Still others believe Xml Schema is too complex and 
> needs to be simplified.
> 
> I am advocating neither of these views.  I believe that profiling Xml 
> Schema will better enable folks to take advantage of "best of breed" 
> tools as they come out, instead of having to wait until each tool 
> implements 100% of the spec.  And let's face it, at best only a few 
> tools makers will go to lengths to support 100% of any spec, let alone 
> Xml Schema.
> Rather than limit oneself to a very few tools, we could create a 
> profile that identifies the most commonly used and easily implemented 
> aspects of the spec.  We could even have a "full" version of schemas 
> and a "tools" version which would be complimentary.
> 
> At any rate, this is all a preface for telling you of our work on a 
> profile for Xml Schema.  We have a fairly extensive one internally and 
> are beginning to document it externally.
> I've created a few blog postings on this as a beginning.
> Some items in our profile are just good practices and others are more 
> of a profile nature.
> 
> Here are the first postings:
> 
> * The case for profiling http://www.hr-xml.org/blog/?p=47
> * No default values http://www.hr-xml.org/blog/?p=26
> * No xsd:union http://www.hr-xml.org/blog/?p=17
> * No xsd:all http://www.hr-xml.org/blog/?p=64
> 
> 
> I hope to post entries on this topic as I am able.  Please let me know 
> what your comments/thoughts/error corrections are.
> Cheers.
> 
> 
> 
> 
> 
> 
> 
> 
>  
> 
> 
>

Received on Wednesday, 19 July 2006 17:22:02 UTC