Re: Issue #3

Francis McCabe wrote:

> 2. The extensibility CSF is supported by XML in a relatively weak sense. 
> Part of the semantics of XML is the default assumption that any tool 
> processing an XML structure that it did not directly understand should 
> ignore that structure. 

Following up the telecon discussion on this...

What I was trying to say is that I'm not convinced that all XML tools 
are supposed to automatically ignore unknown structure. A notable number 
of tools are based on XML Schema in which it is possible to permit an 
element to contain unknown elements but it is also surely possible to 
block this and enumerate all of the elements which may be contained 
within a parent element.

It is possible that my view of the XML world is biased by mostly seeing 
things intended for use with SOAP which has stronger closed world 
assumptions. It is also perfectly possible I'm simply mistaken - my 
experience with XML is limited and not happy.

I certainly don't think XML opposes the extensibility CSF, it can be 
arranged that it permits extension. Though documents like [*] seem to 
suggest that it takes non-trivial thought to sow the seeds of forward 
compatibility into your schema.

...

On reflection perhaps all I'm doing is registering surprise rather than 
raising a real objection. Having registered that surprise I'll now try 
to get over it, and withdraw my objection - go ahead and put the link in.

Dave

[*] http://types.bu.edu/seminar-documents/Versioning.pdf

Received on Tuesday, 26 September 2006 16:15:18 UTC