OASIS Mailing List ArchivesView the OASIS mailing list archive below
or browse/search using MarkMail.


Help: OASIS Mailing Lists Help | MarkMail Help



   Re: XML Schemas: lost "open content models"?

[ Lists Home | Date Index | Thread Index ]
  • From: Rick JELLIFFE <ricko@allette.com.au>
  • To: xml-dev@xml.org
  • Date: Sat, 26 Feb 2000 22:06:45 +0800

Don Park wrote:

>  My particular dislike for the current crop of schema initiatives
> is that all of them seem to target the syntax layer instead of
> the information model layer.  Current syntax-targeted approaches
> decrease extensibility while increasing fragility of XML-based
> documents.  Also, I find current obsession with document-level
> schemas in conflict with the increasing trend toward compound
> documents/streams.  Micro-schemas which define basic building
> blocks are preferable to macro-schemas which define complete
> document structure.

Then Don should love XML Schemas, since that is exactly its basis.
Almost the only thing in it that has document scope is ID/IDREF; and
even there it also provides an XSLT-inspired key mechanism.

I presume by syntax-related approaches, does Don mean
that sequence is bad?  XML Schemas operate on the information
set not the incoming tokens.  It  departs from a simple
sequence-based grammar approach by:

    * wildcarding the <any /> particle in content models, to allow
various kinds of openness;

    * providing an <all /> content model type (similar to SGML &)
which is position-independent (but note that it is currently only available at

the top-level of a content model: some of us feel that anonymous groups are
to some extent bad information modeling practise--if a group has semantic
significance why shouldnt it be named?);

    * providing element equivalence classes and abstract elements, allowing
you to name the slot in which a required element should go, but defer all
type decisions to the schema of the element that will take advantage of it:
you can say "some element for credit card number should go here, but
I wont provide any type or namespace information for what it will be".
This supports frameworks.  (I.e., an abstract element is a non-terminal.)

Rick Jelliffe

This is xml-dev, the mailing list for XML developers.
To unsubscribe, mailto:majordomo@xml.org&BODY=unsubscribe%20xml-dev
List archives are available at http://xml.org/archives/xml-dev/threads.html


News | XML in Industry | Calendar | XML Registry
Marketplace | Resources | MyXML.org | Sponsors | Privacy Statement

Copyright 2001 XML.org. This site is hosted by OASIS