[XML-DEV Mailing List Archive Home] [By Thread] [By Date] [Recent Entries] [Reply To This Message]

Re: XML Schema considered harmful?

• To: mike@a... (Michael Leditschke)
• Subject: Re: XML Schema considered harmful?
• From: John Cowan <jcowan@r...>
• Date: Wed, 5 Jun 2002 16:08:19 -0400 (EDT)
• Cc: xml-dev@l...
• In-reply-to: <LOBBICBLDJIJHPNJFPOLMENJCGAA.mike@a...> from "Michael Leditschke" at Jun 05, 2002 02:32:53 PM

Michael Leditschke scripsit:

> The level of co-constraint checking offered by RELAX NG is an improvement
> on their absence in XML Schema but is this more than leap-frogging? I would
> also add that the situations where I needed co-constraints were often
> content based, e.g. if element X contains "5", attribute Y should be
> present.
> I may have missed it, but the support in RELAX NG seems, by the nature of
> RELAX NG, purely structural. 

Not entirely: in the case you want, RNG can cope.  Consider this schema:

element x {
	(attribute y {text}, "5")
	| (string - "5")
	}

Valid documents:

	<x y="7">5</x>
	<x>4</x>

Invalid documents (Jing says "bad character content for element):

	<x>5</x>
	<x y="8">4</x>

RNG, like XML, doesn't care about the order of the attributes in the instance.

An alternative schema that permits (but does not require) y to be present
if x is not 5:

element x {
	(attribute y {text}, "5") |
	(attribute y {text}?, string - "5")
	}

-- 
John Cowan <jcowan@r...>     http://www.reutershealth.com
I amar prestar aen, han mathon ne nen,    http://www.ccil.org/~cowan
han mathon ne chae, a han noston ne 'wilith.  --Galadriel, _LOTR:FOTR_

• References:
  • XML Schema considered harmful?
    • From: "Michael Leditschke" <mike@a...>

• Prev by Date: RE: Interesting mailing list & a rare broadside
• Next by Date: Re: Interesting mailing list & a rare broadside
• Previous by thread: Re: XML Schema considered harmful?
• Next by thread: Can RDDL and RDF coexist?
• Index(es):
  • Date
  • Thread