The current exchange on the semantics of the 'Resource Type' element
echoes the kind of frustration and confusion we are experiencing
in applying this attribute for the diverse types of information resources
being described for the Everglades Information Network/Everglades 
Digital Library (primarily a sci-tech domain with emphasis on ecological
and earth science subjects).  

At the risk of running a bit long for an e-mail posting, I'm 
duplicating this part of my DC-5 pre-meeting questionnaire since 
it relates to the issue at hand....

||------ warning! long posting ahead -----------------||

Ideally, we would assign to the 'Resource Type' element two different
types of values, reflecting two distinct differences in semantics:

(1) Resource Type as Genre or Document Type

In our domain, this attribute helps users identify the quality of the
information in the resource by describing its path to publication.
For example, a resource type of 'journal article' or 'conference paper'
indicates a higher level of quality by virtue of the form of peer-review
or editorial review the resource has gone through.  Conversely,
a resource type of 'Letter', 'Errata', or 'Column' says something
about the less rigorous filtering process these forms of communication go
through as they go from author to reader.  Most of the well-established
sci-tech information services provide for a 'document type'
element, either as part of the subject descriptors or as a separate
search field.

In other fields beyond sci-tech, e.g., within Arts & Humanities, it
seems there is also an established practice for attributing 'Genre',
something that may tell the user how a photograph was taken or processed,
or how a piece of prose came into publication (this is not to be confused 
with the use of the term "genre' which connotes an artists' style
or a school of art.) There
are numerous thesaurii in these fields that would provide specific
examples of "Resource Type" or "Genre",
and I have collected several examples to bring to Helsinki as a point of
reference.  In the meantime,  I apologize for this woefully inadequate
characterization of 'Genre' and hope those knowledgeable list subscribers
will feel free to enlighten!

Moving along, then....

(2) The other usage for 'Resource Type' that we here in the lower
wilds of Florida have identified is the concept of 'granularity',
which appears to be more analagous to Dr. Malet's "length of
interaction" or "time to access subject matter".  In essence,
this use of 'Resource Tupe' tells the user something about the
substance/comprehensiveness/depth-of-treatment of the ultimate
source.  Is it a piece of data or an entire searchable database?
Is it a journal (the sum of all issues and articles); just one
issue; one article; or a piece or part of one article (e.g., a figure,
a table, or a particular section? Or perhaps, as described in Dr. Malet's
example from the medical field, a mere abstraction or summary of an 
article?

The question of granularity or "what is the extent of the work?" does
appear to matter to the searcher, who may decide to pursue
an item based on some assessment of whether it is "worth the
bother".  The results of Moen's recent study on the quality
of GILS metadata substantiates this contention that the searcher
does want to know, in the earliest stages of resourse discovery,
what kind of granularity the ultimate digital resource offers.
----------------------------------

As we consider how best to convey both these possible meanings
for 'Resource Type', I look forward to learning how other projects
are making best use of this very important element.

Cheers,  

Gail 

Gail P. Clement
Science/Information Services Librarian and
Project Director, Everglades Information Network & Digital Library
Florida International University Libraries, Miami, Florida  33199
E-mail: [log in to unmask] | Phone: 305/348-3417| FAX: 305/348-3408