Lists Home |
Date Index |
- From: Mike Brown <firstname.lastname@example.org>
- To: "'email@example.com'" <firstname.lastname@example.org>
- Date: Mon, 21 Aug 2000 12:23:41 -0600
> Do you always have a separate XML document to model
> each object?
> Isn't it more common to have several different object classes
> in your model, and typically many object instances per
> document, with relationships both within and across documents?
In this case, the object in question is one single "web site", the contents
of which are represented in part by a visual design related component model
(with many types of objects) according to a schema that I devised to meet
the needs of graphic designers who needed a great deal of flexibility; the
components are reusable but also highly customizable each time they are
Yes, of course there are many different object classes. These are broadly
grouped into categories, each category getting a separate XML document. The
categories are determined by the need to keep certain types of data
separate, so that separate processes can act on and manage them. I have the
site's design structure modeled in one or more files as nested elements of
XML or XHTML, a library of static design data (image URIs, heights & widths,
color hexcodes, font sizes and styles, etc.) in another file, content data
in another file, and metadata and data specific to site generation time in a
fourth file. There are references via shared attributes between the
structure file and the other files, except the content data which is only
referenced by the XSLT code that implements some of the structural
Class definitions are implied or specified through DTDs. Before now, I've
never had a need to represent the class definitions in XML. The "entire web
site" class is thoroughly implied and is not modeled by any element in the
XML files at all, nor by one single XML file (and such a file would never be
in a location stable enough to be referenced by a URL).
What we're doing now is producing national language variations of sites, and
there is a need to link from one site to another. I want to uniformly
express the relationship between the sites, i.e. in that 4th file say "the
site we've modeled in this collection of files is the US English version,
and there also exist these related sites: a Canadian French version and a
Belgian Dutch version" (and yes, I realize they'll get out of sync.. that's
not a concern right now though).
> Expressing relationships (UML: associations) between objects
> even within one XML document is not that simple, because XML
> gives you so many ways to do it.
Exactly, and the examples you gave are all approaches that I am currently
using. Before I go and continue down that road, I want to know what more
formal options I have.
I received some helpful info in private email, too, so here is a summary of
all the responses I received.
1. XMI: XML Metadata Interchange (omg.org)
2. RDF Schemas
3. SOAP serialization (sec. 5 of SOAP 1.1 note)
4. The Castor Project (http://castor.exolab.org/)
Mike J. Brown, software engineer at My XML/XSL resources:
webb.net in Denver, Colorado, USA http://www.skew.org/xml/