ejfreed at infocanvas.com
Wed Jun 16 22:48:50 BST 1999
The exact terms are ones that reasonable people can (and do) disagree on. So
these terms are
not advertised as authoritative, but to me one perspective is as follows:
the data environment has no idea what its own meta-data is
the data environment can somehow get meta-data information
The meta-data for a system is modeled in its own data (ie a class is an
The meta-data for a system is dynamically modeled in its own data (it can
change at run-time)
For instance you could add a completely different type of semantic property
to a type, and have
meta-data be based on it.
The meta-data for a system is dynamically modeled in its own data and boots
from this definition.
Various systems over the years have achieved various levels in various ways.
There are other possible
levels for other combinations of capabilities.
Java for instance could be claimed to be level 2 (except no-one writes java
and perhaps level 3 if you allow the system to compile and load new classes.
(that would be pushing it)
Level 3 and 4 require interesting chicken and egg boot procedures.
I really do not know where this sort of thing is discussed except implicitly
and explicitly in
the proceedings of special interest groups such as ACM and IEEE. There are
probably books out there
but I cannot recommend any specifically.
XML is interesting in that since it is very dynamic in both its meta-data
and data dimensions, it has the
possibility of supporting higher level more open and extendable systems.
The XML system I am writing will be level 4. It will boot its first class
object definitions, from an XML-SCHEMA definition,
with the schema reader being essentially a bootstrap program. At run-time
however the model can be
changed and extended.
From: G. David Kuhlman [mailto:dkuhlman at netcom.com]
Sent: Wednesday, June 16, 1999 1:17 PM
To: ejfreed at infocanvas.com
Subject: Re: X-Schema concerns
> The latest arguments regarding the usefulness of XML schema languages
> DTDs are the most interesting. I have long used what are sometimes called
> 'reflective' object models, where the 'meta-data' can be thought of as a
> specialization of 'data'. This unified approach in my experience is more
> than a conceptual nicety, it provides long term implementation advantages.
> single model for creation, storage, queries, fetches, modification,
> browsing, editing, translating, etc means that your large scale mature
> system is simpler to understand and has much less implementation code in
> For instance, the reflective system I have most recently constructed, has
> browsing tool, where with only one tree view model, can browse through
> meta-data and data 'boundaries'. This unified approach, I find, pays off
> immediately in very practical terms.
I'm interested in understanding what you call the "'reflective'
object models". Can you suggest a resource or reference where I
could learn more about it? Even hints and suggestions would be
The application I work on edits objects that have an XML
representation (along with others). I'm guessing that our
application would benefit from the ability to deal with objects
that are self describing and that help tell the object editor how
to display and modify them. Am I a little bit on track here?
In the future, when more and more of the objects in our
applications have an XML representation, then the ability of
objects to describe themselves and the ability to annotate an
XSchema with information that helps an editor (and possibly other
tools as well) will become more helpful and useful.
Thanks for any pointers and help.
xml-dev: A list for W3C XML Developers. To post, mailto:xml-dev at ic.ac.uk
Archived as: http://www.lists.ic.ac.uk/hypermail/xml-dev/ and on CD-ROM/ISBN 981-02-3594-1
To (un)subscribe, mailto:majordomo at ic.ac.uk the following message;
To subscribe to the digests, mailto:majordomo at ic.ac.uk the following message;
List coordinator, Henry Rzepa (mailto:rzepa at ic.ac.uk)
More information about the Xml-dev