Whilst roaming about in lala-land today, I stumbled on this site:
http://www.cetis.ac.uk/content/20030127164729
"The one standard, LOM and the semantic web" (01)
The page is a response to a paper by Stephen Downs, in which a challenge is
made to the need for just one Learning Object Metadata standard. (02)
The argument, if I may summarize it by plagiarizing from the Web is this:
"Objects are best described using multiple vocabularies. There is no way to
determine which vocabulary will be relevant to either an author or a user
of a given object. Trying to stipulate a canonical vocabulary a prior
needlessly reduces the effectiveness of a system of communication." (03)
Iben Browning used to opine that, given a clean sheet of paper, that paper
was available to express anything you could think of, but, the minute you
began to draw lines on that paper, the range of things it could express was
dramatically reduced. (04)
As I was reading both the challenge (it's linked on that page) and the
response, I was thinking about the Entity-Attribute scheme cooked up by
Gregory Rawlins and his colleagues and implemented in the Knownspace
project. It seems to me that the E-A scheme, in no way, presents a
limitation on how Entities are characterized; you are more than free to
shovel at an Entity any Attribute you can cook up without regard to any a
priori ontology. (05)
I know, I'm rambling here, but I am reminded not only of the Knownspace
approach, but the topic maps approach, each taking a minimalist ontological
commitment with which entire universes of discourse can be represented. RDF
is similar, with a minimalist ontological commitment on which RDFS can
layer further ontological commitments according to user needs. (06)
What seems to be argued about in the cited URL is an application, one in
which the nature of the ontological commitment made to learning objects is
held in question. I don't think I'm smart enough to take a side in such an
argument (read: I'm always interested in the views of others in such
matters), but I am fascinated with the nature of the argument; I tend to
reframe the argument as any true XTM-wonk would: are we not arguing about
how to represent properties of subjects? (07)
As my friend and partner Sam Hunting is fond of pointing out, "Topic maps
have the objective of making sure that everything that is known about a
subject is accessible from one place." (08)
I am, in some sense, ignoring an underlying thesis in the argument
mentioned above: the need for ontological commitment as expressed in
metadata standards. On that, I would argue that such a need may be reduced
by the introduction of an appropriate topic map that captures properties of
subjects in such a way that all users and authors are satisfied. (09)
The reply seems to agree with Downes, seems to sit on the fence regarding
the need for multiple vocabularies, but essentially supports RDF as the
proper vehicle for the job. Here, I would argue that RDF, when combined
with the topic maps Standard Application Model, which facilitates merging
of maps, would be better suited to the job. For me, it's interesting to
note that the topic maps application model never finds its way into such
discussions. What starts out to be a discussion about metadata is, at least
for me (while playing Liz Story in the background), a discovery that,
perhaps, the right questions are not being asked. (010)
Running out of EUROs to toss around...
Cheers
Jack (011)
---------------------------------------------------------------------------
XML Topic Maps: Creating and Using Topic Maps for the Web.
Addison-Wesley. Jack Park, Editor. Sam Hunting, Technical Editor (012)
Build smarter kids globally to reduce the need for smarter bombs. (013)
--
This message is archived at: (014)
http://collab.blueoxen.net/forums/cgi-bin/mesg.cgi?a=tools-yak&i=5.1.1.6.0.20030315154152.028ea3e8@thinkalong.com (015)
|