--- Begin Forwarded Message ---
Date: Fri, 2 Apr 1999 10:51:41 -0500
From: David Allen <[log in to unmask]>
Subject: Re.: CORC and the Dublin Core
Sender: David Allen <[log in to unmask]>
 
 
 
 
 
     This is a follow-up to my previous message on CORC and the Dublin Core, and
also a contribution to our ongoing thread about indexing digital images.
 
     When I posted my previous message about the CORC project, I did not realize
that there is already a publicly available CORC database, and that it contains
quite a few records for digital maps and other map resources.  I would strongly
suggest that anyone interested in indexing digital images of maps take a close
look at this database.
 
     The URL for CORC is:
<http://www.oclc.org/oclc/research/projects/corc/index.htm>.  Once you get to
the home page, look for "Using CORC" around the middle of the page, and click on
"CORC System."  When the next page comes up, click on the "Go!" button (ignore
the user ID and password box).  On the page which comes up after that, look
under "Search" and click on "Catalog."
 
     When I typed in the keyword "maps," I got hits for 4,310 records.  It is
possible to do complex boolean searches on various fields.  Records can be
displayed in both Dublin Core and MARC formats.  The records retrieved are a
real hodge-podge: digital images of maps, regular cataloging records for paper
maps, books about maps, and various Internet sites that are in one way or
another concerned with maps.  When the record contains a URL for a site, a
window opens up and connects you with the site.
 
     For me this is a very exciting development, and already a useful tool for
searching out cartographic information.  However, it has some limitations, and
needs further refinement.
 
     One problem is that there seems to be no way to call up just digital
images.  A search for maps of a particular place will pull up records for both
digital and non-digital materials.  Also, searches will call up records for both
whole collections and for individual items.  Unless you are doing a very
specific search, such as an author or title search, there is no way to restrict
your search to just records for individual maps.
 
     I am sure that the above problems can be dealt with fairly easily by better
indexing.  However, the existence of a database like this raises a whole host of
questions and problems for those of us who are interested in indexing individual
images of maps available on the Web, or in the "brief cataloging" of maps using
the Dublin Core.  The problems are not new, but it is becoming more urgent that
we solve them.  I will list the problems that I am aware of, and suggest some
possible solutions.
 
1.  What is the MINIMUM amount of information that should be required for a
record in Dublin Core format if it is to go into a database like CORC?  (I would
opt for author, title, size, date of the original map, URL for the digital
image, any restrictions on copyright or use.)
 
2.  What information beyond the minimum is DESIRABLE for such records.  (My list
would include scale, keywords or subject headings, notes, longitude-latitude,
resolution of the digital image, image type).
 
3.  What information is PERMISSIBLE in such records?  Many librarians and others
would doubtless want to include detailed information about the digital record,
such as is found in FGDC metadata, or even graphic indexes.  How much of this
should be included, or is it better just to provide links to sites that have
this type of information?
 
4.  Should these records include holdings information for maps in paper form?
If so, how should this information be represented?
 
5.  What do you do when one library holds the paper map, and another has posted
a digital image of it?  (This is the case at Stony Brook, and at least at
several other libraries.)
 
6.  What do you do when there are multiple images on the Web of several
different paper copies of a single map?  (This is already the case with Blaeu,
Nova Belgica, 1635.)
 
     We really do need a group of people to address these questions and
doubtless other questions as well.  At this point, it looks as though this could
best be done within the context of the CORC project.  Therefore, I would urge
all Mapsters who are in a position to do so, to encourage your institutions to
participate in the CORC project, and to get involved in it yourselves.  See the
CORC home page for information on how to participate.
 
--- End Forwarded Message ---