Well, Mike, here I am trying to get my socks out of the dryer
and packed so I can catch a plane tomorrow to Melbourne, and
don't have a lot of time ... but, let me give this a try. You start
with this idea of "core" and "extended" metadata. New words
for very old concepts. Scholars of librarianship (read, e.g., Patrick
Wilson, Lubetzky, Svenonious, etc. etc.) have been exploring this
distinction for most of the last century and a half. The first
is supposedly objective (i.e., who is the author and what is the
title) and the second is subjective (i.e., what audience is this for
etc. etc.). The first is supposedly useful in resource discovery and
retrieval and the second is fraught with danger.
I am coming to Melbourne and leaving behind 60 students studing
the foundations of cataloging (in the traditional sense of libraries
as we have known them). We are in the midst of dealing with the
most SUBJECTIVE of all aspects of that process--denoting the
intellectual content of the entities being described, determining
SUBJECT access. So, I assume given the dangers of the subjective,
that only derived keywords (as in taken directly from the text)
should be acceptable in DC.Subject? _Maybe_ OK for
text; pretty rough for nontextual resources. Use of such a thing as a
scheme like Library of Congress Subject Headings to assign
a subject (can't get more subjective) is taboo because it is not
directly and objectively derived from the resource?
Mike, I, too, look at the matter in terms of the objective
and the subjective. I think that discovery depends on both.
In traditional cataloging (as in libraries), we speak of descriptive
cataloging (I think that is somewhat coextensive with your "core")
and the providing of "intellectual access" (which I think is _somewhat_
coextensive with your extended) which is always subjective. Now,
saying that the audience for an educational resource is the "2nd
grade" can be either descriptive (core) or subjective (extended). Let's
look at it in the descriptive/objective sense first. Suppose, Mike, that
we are looking at a resource in which the creator says: "This is designed
for use in teaching simple addition to first graders." If, in creating
metadata describing the resource, I note (in a structured way) that this
resource is on the subject of MATHEMATICS with an _intended_
audience of GRADE 2, have I not _described_ some aspects
of the resource? Note that the first (mathematics) is subjective (and,
supposedly, OK) and the second (grade 2) is objective--it directly
describes an objective aspect of the resource (and not OK).
Now, a determination that a resource is useful for an audience of
second grade students could be highly subjective. Imagine a room
full of munchkins sitting around creating metadata for educational
resources. Imagine further that the resources say nothing about
audience (or subject!). These munchkins select subject terms from
a controlled vocabulary _AND_ make subjective determinations
regarding audience. We have lived with the former kind of determination
for a LONG TIME! We assign subject(IVE) vocabulary terms from
the Library of Congress Subject Headings, the Art and Architecture
Thesaurus, etc. and have "parked" physical items on library shelves
with such subject(IVE) schemes as the Dewey Decimal and Library
of Congress classification schedules. I find it ironic that we now
want to explore an online universe that is infinitely larger and
potentially more complex in which we reduce discovery to
the title and author elements of a resource (and perhaps some
mish mash of uncontrolled keywords--objective as they might be).
Now, while I personally would consider this second (highly subjective)
use of an audience element objectionable, I am no more
willing to throw out the idea of it than I am to throw out DC.Subject.
So, I have no problem with an AUDIENCE element and a best
practice that says it should be used for the _descriptive_ aspects
(as in, when a resource SAYS it is for a specific AUDIENCE).
Duly noting such a fact in a structured way is just as objective as
transcribing the name of the author or the title of the work. If we
cannot tolerate the occasional misuse, then I challenge on the
very same grounds our willingness to accept anything resembling
DC.SUBJECT beyond simple keyword extraction (nothing so
subjective as adding subjects terms from a controlled vocabulary).
Mike, I have come to the conclusion that you are dead wrong
when you assert that "there is no pressing need for further extended
elements and that the use of Comments/Annotations allows
greater flexibility and opportunities for value add." It's an assertion
that abandons the very principle resting under structured
metadata. While we could say everything we need to say about
educational resources in an "annotation," discovery based on
structured metadata would not be possible. If this is truly the
case, then why don't we achieve more "flexibility and and
opportunities for value added" by pulling DC back to a single
element--Comment/Annotation? I think it is because we would
be right back with the same world we had that DC was supposed
to help mend.
Finally, we talk a great deal in this community about "discovery."
DC came up with title and creator etc. because there was a history
of discovery based on those elements in information seeking behavior.
They didn't just spring from nowhere. In our early work, we looked at
hundreds of thousands of archived inquiries in the AskERIC database
to see what USERS were asking for ... how they were trying to discover
resources. These aspects you'd dump into an undifferentiated annotation
are the very sorts of information they used to try and discover resources.
Stuart
----------------------------------------------------------------------------
------------
Stuart A. Sutton (206) 685-6618 (V)
University of Washington (206) 616-3152 (F)
School of Library and Information Science
Box 352930
Seattle, WA 98195-2930 [log in to unmask]
GEM http://geminfo.org (Project)
http://www.TheGateway.org
----------------------------------------------------------------------------
------------
-----Original Message-----
From: Mike Currie [mailto:[log in to unmask]]
Sent: Tuesday, February 15, 2000 7:27 PM
To: DC-Education (E-mail)
Subject: Use of Comments v DC.Pedagogy, etc.
Hi everyone,
I've been monitoring the list since it started but it was meeting Stu
and Liddy last Monday in Sydney that finally inspired me to share a few
thoughts that arose out of that meeting.
I apologise for the length of the posting.
1. Currently there seem to be two schools of thought. There are those
who want to add value to metadata in order to "give teachers what they
are looking for". These argue that adding educational elements such as
DC.Pedagogy gives the resource a context in which to use it, ie. how
could we best use the resource.
The other group argues that metadata should describe the resource, not
the context, ie. it should identify the nature of the resource per se
not how it could be used. This is Stu's and DCMI's position. They argue
that the use of context metadata is capable of misinterpretation and
ongoing argument re the use or value of the resource. DCMI argues that
metadata should be value free.
I believe the discussion then boils down to 'What is the role of
metadata?'
2. I suggest that metadata for a resource can be grouped into core and
extended metadata. Core are those elements that describe the basic
attributes of a resource (eg. title, creator). These elements are by
their nature objective. They have no relationship to how or for whom the
resource is used, ie. the context. Other elements such as the proposed
DC.Pedagogy or DC. Audience are conversely essentially subjective. They
derive from one persons opoinion as to the particular application or
context of a resource. Who defines what the particular pedagogy of a
resource is? If it is a NEF (non-educationally focussed) resource, the
Creator (even if they make the effort to add metadata) is unlikely to.
If it is indexed by a third party, who is to say that their view is
authoritative. However, assuming the a tag is then added, the resource
from then on becomes
'tarred with that brush' (to use an expression). For right or wrong,
people will judge the value/application of that resource according to
the available metadata.
My first argument then is that the application of subjective (or context
based) metadata is potentially fraught with danger.
3. I suggest that metadata has two functions: resource discovery (ie
locating suitable resources) and resource description (providing useful
value-added information about the resource).
The metadata for most resources will include the core items. However
given that all elements in DCMI are optional, extended elements will
only be used sparingly (where a particular indexer is enthusiastic
enough to include elements which he/she thinks is particularly
relevant). Hence only core elements are really useful for resource
discovery across a wide range of resources. It is far more logical to
search for a resource by title or keyword than as an example of a
resource that will keep a class occupied for 35 minutes.
4. Having discovered a potentially suitable resource using core
elements, it is reasonable to allow the searcher to find out more about
the resource. This could be done using extended elements (DC.Pedagogy)
but alternatively it could also be done by the judicious use of free
text Comments or Annotations (see Albert Ip's work on the use of
Annotations as Type 1 data) saved as separate documents and linked to
the original resource by means of an element such as DC.Relation. The
technical solution for this might be achieved using a bi-directional
link to a link table.
The use of Comments helps to retain the integrity of the metadata as a
vehicle for resource description and
allows ongoing value adding to the resource by practitioners. Most of
these would not have rights to alter the original metadata record.
5. Hence I come to the conclusion that there is no pressing need for
further extended elements and that the use of Comments/Annotations
allows greater flexibility and opportunities for value add.
Regards
Mike
--
Michael Currie
EdNA Project Officer (Higher Ed.) Ph. +61 3 9344 9578
University of Melbourne Fax +61 3 9347 9106
Parkville, Victoria, Australia Email
[log in to unmask]
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|