Jordan's email of yesterday on this topic once again proves
the theory of simultaneous invention, plus, of course, that
great minds think alike! I hope there is time in Helsinki to
address this matter.
I realize and appreciate all the thought and work that has
gone into the ResourceType draft, and it seems to me that we still need to work
on this more. (Yes, I'm willing to help, not just stand off in a corner and make
rude remarks :-)
I. The September 23, 1997, "Dublin Core Metadata Element Set:
Reference Description" defines Resource Type as, "The category of the resource,
such as home page, novel, poem, working paper, technical report, essay,
dictionary." It refers the reader to
http://sunsite.berkeley.edu/Metadata/types.html for the draft.
The "Dublin Core Resource Types : Structuralist Draft, July 24, 1997"
(printed out here September 25, 1997) states in the first full paragraph,
second line, that, "This element is to describe the genre of the item being
described," and lists the top-level Resource Types as: text; image; sound;
software; data; and interactive.
"Genre" is often taken to mean referring to the intellectual content of the
item, as contrasted to its physical form. For example, "novel" is a genre term
in that a novel can appear printed on a piece of paper, on microform, or in
digital form on someone's computer screen - it is still a novel, and only the
physical form (the carrier of the information) has changed.
***Is this what it is intended to mean here?*** This key point needs to be
decided before we proceed.
II. I have held off on sending this email because thesaurus-building is not my
specialty; I just did not feel comfortable about the list, and having just gone
through a substantial amount of committee work in building a list of form/genre
terms for geospatial information, I figured that I knew only enough to hunt up
someone who specializes in this work before I said anything. I have the good
fortune to work with a person - Linda Hill ([log in to unmask])
- who has worked with indexes, abstracts, and thesauri for some years. After
looking at the list of Resource Types in the Structuralist Draft, she has
contributed the following points, plus an appendix that is explained later in
this section.
If we are going to develop a short list of terms that will assist persons, it
would be a good idea to build it following the thesaurus guidelines available
from ANSI/NISO and ISO. Some of the following specific principles can be
found there.
a. Resource Type terms should be nouns, not adjectives; e.g., "interactive"
would need to be changed to "interactive items" or something like that.
b. Terms should stand on their own and be understandable when they are
used as descriptors - that is, their meaning should not be dependent on
seeing them within the hierarchy of the scheme. Therefore, a term like
"executable" under "software" should instead be "executable software".
c. The choice of singular vs. plural forms of the terms is debatable and
either could be used. However, the ANSI/NISO Z39.19 standard calls for
"count nouns" to be expressed as plurals ("names of objects or concepts
that are subject to the question 'how many?' but not 'how much?'). Also,
if the primary purpose of the Resource Type terms is for searching, seachers
are more likely to think in terms of plurals rather then singular. It is
generally only at the cataloging time when there is one object to be
described that the singular term is the most likely form to use.
d. There are shallow and deep hierarchies, and arguments on both sides as to
which is better. The limited number of top-level nodes and an orderly grouping
under them are more effective tools for use by the general
public.
e. The use of cross references (1) from nearly synonymous terms for the
same sort of thing and (2) between related terms in the thesaurus of terms
is essential for a useable tool.
f. The relationship between a broad term and a narrow term should be kept
strictly an "is a" relationship - that is, a genus/species relationship. It
should always be true that a term "is" also truely its broader term - so that
if the broader term were also apply as a descriptor to an item, it would
also be true.
g. It is difficult to keep the Resource Type list free of the closely
related set of terms that describe how an item IS AVAILABLE. A "homepage" is a
method of presentation and does not belong in this list. What the Web page has
on it - what it IS - is what would be described by this set of terms.
The cataloging world has found that keeping genre and carrier terms in separate
lists is extremely difficult, and has in effect shoved them together in a USMARC
field (REBECCA - have I correctly presented this?). If we decide to have both
genre and carrier terms in one list, fine; but we need to define ResourceType to
cover both.
h. The notion of creating stand-alone terms by concatenating the hierarchy
(e.g., "image.interactive") is not a good idea because (1) the resulting
terms are not familiar and (2) it forces users to know the hierarchical
structure of the terminology.
i. With a hierarchical structure of terms, it is not necessary to list
every possible genre form because: (1) if a specific term is not listed for
an item, a term that is higher in the tree can be used instead; and
(2) thesauri are expandable - more terms can be added for local purposes or
the 'main' authoritative set can be expanded through time, as long as the
basic structure accommodates all anticipated categories of terms.
What follows in an appendix is the hierarchical listing of a draft Resource Type
thesaurus Linda Hill has been working on. It is not complete; there is enough of
it to give a general outline. This listing shows only the valid terms and their
hierarchical relationships - not the invalid (synonymous terms or the Related
Term links). Linda would be happy to put the draft thesaurus up as an html
document so we DC persons can explore all of the terms. If this approach to
structure is accepted, the details of the terms and relationships can be the
focus of discussion.
III. As I noted above, what led me to talk with Linda Hill about this matter
was that I could tell that a great deal of thought and hard work had gone into
the "Structuralist Draft," and yet from the point of view of spatial data and my
years of experience of cataloging, there were points that did not seem to
follow. Some specific examples follow:
- Almost everything is "data" at some level, so while I see what
is being aimed toward, this is not a good category.
- The list of "text" items does not all seem to be at the same level, nor does
it list all possibilities at a level. If one has "dictionaries," then one must
also have "encyclopedias," and so on.
- In compiling a list like this, it would be
extremely helpful to have a printout of the Library of Congress's list of what
are called free-floating subdivisions for subject headings (REBECCA, please
correct me if I have that name wrong) - it's a list of subdivisions (e.g.,
"Maps," and so on) that may be used with almost any thematic subject (e.g.,
"Geology", "Housing", etc.). My geospatial-data form/genre group used this,
and found it saved us time and assisted us in constructing a logical structure.
- Strictly from the viewpoint of spatial data - this list creates 2 possible
places to put spatial data, since large amounts of it is remote-sensing images
(e.g., aerial photographs) - yet "spatial data" is listed in only one location.
Mary Larsgaard
Alexandria Digital Library/Davidson Library
University of California
Santa Barbara
------
APPENDIX
RESOURCE TYPE TERMS - DRAFT THESAURUS FROM ALEXANDRIA DIGITAL LIBRARY (L. Hill)
cartographic works
. atlases
. maps
collections
. archives
. catalogs
datasets
. survey data
events
images
. art works
. . art originals
. . art prints
. . art reproductions
. graphic materials
. motion pictures
. photographs
. . aerial photographs
. . photographs from space
. remote-sensing images
. . aerial photographs
individuals
. experts
multimedia packages
musical works
organizations
projects
recordings
. sound recordings
. video recordings
services
. interactive services
. . chat services
. . discussion lists
. . games
. . news groups
. reference services
. . cataloging rules
. . classification systems
. . dictionaries
. . directories
. . gazetteers (place names dictionaries)
. . metadata standards
. . subject heading lists
. . thesauri
software packages
textual works
. abstracts
. articles
. bibliographies
. biographies
. . autobiographies
. conference papers
. correspondence
. fictional works
. . dramatic works
. . poems
. monographs
. . monographic series
. proceedings
. reports
. serial works
. . journals
. . magazines
. . newsletters
. . newspapers
. theses
