RE: XSD feature check-lists

I think there are a number of processors that are very close to complete
conformance, and it's very hard to capture their remaining restrictions on a
simple tick-list. For example, one of the few remaining restrictions in
Saxon is:

"Default or fixed values specified for an element whose type is a complex
type with complex content are ignored when validating instance documents."

http://www.saxonica.com/documentation/conformance/schema.html

You'd need a pretty long feature list to capture things at that level of
granularity.

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


> 
> Dear Noah,
> 
> While I see your point, I feel that XML schema is about 10 to 
> 20 times more 
> complex than an XML parser (a rough estimate based on implementation 
> experience).  Therefore, it isn't necessarily appropriate to 
> extrapolate 
> what has worked well for XML to XML schema.
> 
> I also wouldn't expect vendors to fill in a feature profile 
> and never work 
> on their code again.  Quite the contrary - I can see that a 
> feature matrix 
> could be quite a battle ground and motivator for vendors to implement 
> features so that they can out do their competition.  This 
> would ratchet up 
> the level of support among vendors.
> 
> The result would be that customers would have a clear picture 
> of what is 
> available, and vendors would know where they need to do 
> better.  If feature 
> support wasn't 100% after a while, standards people could go on about 
> vendors not properly implementing standards and add pressure 
> based on hard 
> evidence!  Looks Win-Win-Win to me!
> 
> Regards,
> 
> Pete.
> ____________________
> Original Message From: <noah_mendelsohn@us.ibm.com>
> 
> > Pete Cordell writes:
> >
> >> I don't know whether this has been suggested
> >> before, but as an alternative to defining domain
> >> specific profiles of XSD, maybe the W3C can define
> >> an official, standard checklist of features that
> >> vendors can fill-in and display on their web
> >> sites.  This would give customers an equal basis
> >> for assessing products, and give vendors guidance
> >> on where they can improve their products.
> >
> > I think that encouraging vendors to be honest about the 
> status of their
> > products is a good thing, but appearing to encourage 
> subsets is not.  How
> > would we feel if for XML itself we suggested a checklist 
> along the lines
> > of:
> >
> > _ do you support attributes?
> >
> > _ do you support entities?
> >
> > _ do you support the short form of empty tags?
> >
> > _ do you handle whitespace correctly?
> >
> >
> > etc.  XML interoperability would be seriously damaged.  The 
> point of XML
> > is that you don't get a choice about these things and we 
> don't encourage
> > reporting of subsets.  We tell you not to claim you're 
> XML-compatible
> > until you do all of these things.  As a result, customers 
> find that XML
> > interoperability is pretty good.
> >
> > The same concern applies to schema.  While a few vendors 
> are still having
> > particular trouble and that is causing unfortunate confusion, my
> > impression is that a number of the more careful XML schema 
> validators are
> > showing quite good conformance and interoperability.  Of 
> course, there are
> > some things like data binding tools that aren't even trying 
> to support the
> > whole language, though some of the things they subset are in XML not
> > schema (e.g. mixed content.)
> >
> > Bottom line: while schema does have well known complexity 
> issues, I think
> > we should be very reluctant to promote ad-hoc subsetting in 
> particular
> > implementations.  I'm afraid that publishing a matrix in 
> which you can
> > document your non-conformance may appear to do just that.
> >
> > If you're talking about things like databinding tools, I 
> support what I
> > took to be the consensus of the recent schema workshop:  
> tools should
> > support all of schema and XML, but should optimize the 
> subsets most likely
> > to be heavily used and/or a good match to the environment.  
> A checklist to
> > document such a subset may make sense, and I think that as 
> a result of the
> > workshop the W3C is considering a closer look at that area.
> >
> > --------------------------------------
> > Noah Mendelsohn
> > IBM Corporation
> > One Rogers Street
> > Cambridge, MA 02142
> > 1-617-693-4036
> > --------------------------------------
> >
> 
> 
> --
> =============================================
> Pete Cordell
> Tech-Know-Ware Ltd
>                          for XML to C++ data binding visit
>                          http://www.tech-know-ware.com/lmx
>                          (or http://www.xml2cpp.com)
> =============================================
> 
> 
>

Received on Thursday, 4 August 2005 10:22:52 UTC