Short Essay: Squeezing RDF into a Java Object Model

[W. E. Perry]

David Megginson wrote:

> Obviously, there's a much more complex model underlying RDF than the
> spec lets on, and that model affects not only the ease or difficulty
> of implementing an object model, but also the difficult of many
> standard operations like queries against a collection of RDF
> statements and storage in a relational database.

May I respectfully submit that the problem is not the complexity of the model underlying RDF,
but its simplicity and relative freedom from restriction, permitting the very sort of
extension that leads to the implementational problems David Megginson illustrates. Or, stated
from the opposite perspective, obvious real-world implementations of RDF build upon the
assumption that the standard specifies, or at least implies, a more holistic view of metadata
than it does. The unfortunate truth is that implementations of RDF--even as they grow
step-by-step as complex as David Megginson illustrates--never model anything more than
individual resources:  they do not, even as a by-product, model the body of modeled resources
as a whole.

Effectively it is just such a cumulative body of metadata which David Megginson is seeking. It
would provide the ability to refer to components larger than simple subjects, predicates and
objects, such as resources as objects or prefixes as subjects. Such a framework would
comprehend and permit reference to any such objects either top-down, from the perspectives of
their larger containers, or bottom-up from the perspective of their sub-components.

In fact, it appears that this inability to implement the innate human assumption of a larger
framework, not specified by the simple details of RDF, is a shortcoming not just of RDF but of
structured markup generally. Recent debates on this list about the unification of XSL and
XLink revolve around symptoms of the same problem. The increasingly centrifugal nature of the
whole body of XML standards could reasonably be described as the failure of building upon the
details to produce as a by-product a framework which multiplies the interconnections and the
interdependence among them.

My own work is in implementing a database engine which operates directly upon XML markup. Its
first premise is that XML markup describes, primarily, structure and that manipulating XML
documents on their own terms means managing them on the terms of whatever structure the
instance markup describes. From my (hardly disinterested) perspective, it is this ability
which David Megginson seems to be wishing for in describing the implementational difficulties
of RDF:  if, for example, an object is a resource, it should be manipulable as either (and
both) an object and a resource. Implementation of the simple details of RDF will not provide
that, but database tools which can operate on structure described by XML markup, to whatever
level of complexity, can. As David Megginson's examples illustrate, the implementation
strategy implied by RDF (and other XML specification) details is agglutinative, an inherently
linear process. We need implementation tools--and not just for RDF--which are also
agglomerative, building a larger ball or sphere of interconnected structure from the details
of instance markup as it is processed.

Walter Perry



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;
(un)subscribe xml-dev
To subscribe to the digests, mailto:majordomo at ic.ac.uk the following message;
subscribe xml-dev-digest
List coordinator, Henry Rzepa (mailto:rzepa at ic.ac.uk)