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

Re: Non-deterministic content models

• To: Gustaf Liljegren <gustaf.liljegren@b...>
• Subject: Re: Non-deterministic content models
• From: John Cowan <cowan@m...>
• Date: Tue, 15 Jul 2003 00:05:39 -0400
• Cc: xml-dev@l...
• In-reply-to: <3.0.6.32.20030715055242.01069350@p...>
• References: <3.0.6.32.20030715055242.01069350@p...>
• User-agent: Mutt/1.3.28i

Gustaf Liljegren scripsit:

> This wording makes me wonder if there is any case where a non-deterministic
> content model would acctually be useful? That is, is the restriction a
> compromise, or can a non-deterministic content model always be simplified
> (like many algebraic expressions) to something shorter with the same meaning?

Non-deterministic models cannot always be simplified, and often it's not
convenient to do so.  For example, the model

	foo, foo, foo?, foo?, foo?

is non-deterministic but clear; the deterministic version is

	foo, foo, (foo, (foo, foo?)?)?

and this is a very simple case.  In more complex cases, the conversion
is a big pain:  consider

	(foo, foo, foo, bar, bar?) | (foo, foo, bar, bar?, bar?)

RELAX NG supports non-deterministic models freely.
See http://www.relaxng.org .

-- 
Do what you will,                       John Cowan
   this Life's a Fiction                jcowan@r...
And is made up of                       http://www.reutershealth.com
   Contradiction.  --William Blake      http://www.ccil.org/~cowan

• Follow-Ups:
  • Re: Non-deterministic content models
    • From: Eric van der Vlist <vdv@d...>

• References:
  • Non-deterministic content models
    • From: Gustaf Liljegren <gustaf.liljegren@b...>

• Prev by Date: RE: Good Times Ahead for "Sharecroppers"? (Was: [OT] Tim Bray on Slashdot)
• Next by Date: Re: Non-deterministic content models
• Previous by thread: Non-deterministic content models
• Next by thread: Re: Non-deterministic content models
• Index(es):
  • Date
  • Thread