----------------------------Original message---------------------------- State University of New York at Stony Brook Stony Brook, NY 11794-3331 David Y Allen Library-Reference 516 632-7110 31-Jul-1995 10:08am EDT FROM: DYALLEN TO: Remote Addressee ( [log in to unmask] ) TO: Remote Addressee ( [log in to unmask] ) Subject: Standards/Guidelines for Digital Preservation I would like to respond to several of the points raised in the recent spate of e-mail messages on digital preservation of maps. First, I would second the motion that anyone seriously interested in this subject download and read the final report of the Columbia oversized color images project. This report addresses most of the issues that one must understand to be an informed participant in any discussion of this subject. The report is available from http://www.columbia.edu/imaging/html/largemaps/ Second, I think the whole discussion points up the need for authoritative standards or guidelines for digital imaging of maps. The issues are complex enough to be daunting to anyone attempting to go it alone, but it is not difficult in principle to develop formulas to determine what dpi/pixels/resolution/file size is necessary to produce images of various degrees of quality for different purposes. My main concern is to achieve some degree of consensus as to what resolution is necessary for the creation of "research level" images (both color and black and white). A committee made up of librarians, researchers, and computer specialists should be able to draw up a set of standards for the rest of us to apply fairly easily. This might be a good project for one of our professional organizations. I can't resist commenting on a few secondary issues. I think that the debate about whether digitization is a form of preservation is something of a non-issue. I trust that nobody would seriously advocate throwing away the original after a map has been digitized. As of now, archival digitization projects dealing with large maps have to work with 4 x 5" color transparencies, and these transparencies capture much more information than can, at least at present, be preserved in digital form. Although I have not seen any studies on the subject, some references in the photographic literature make me reasonably certain that a file would have to be more than a gigabyte in size if every silver halide particle in a color 4 x 5 were to be translated into pixels. The moral: digitization projects should rely on archival color negatives as the primary means of preservation. Ease of access and the ease with which they can be copied makes digital images valuable as a secondary means of preservation--by cutting down on the use of both the originals and archival copies of color transparencies. The most important question that needs to be solved is what level of resolution is "good enough" for MOST research purposes. I have considerable sympathy with the concern of Tom Neff and others that digital projects capture every conceivable detail of a map that might be of interest to researchers now or sometime in the future. But questions of expense, file size, and even what is possible with present technology cannot be dismissed so easily. If we hold out for this high a standard, we will have to forget about digitizing anything until such (hypothetical) time as technology improves enough to make it possible. Researchers interested in this level of detail will want to consult the originals anyway, and we should not let paranoia about digital images "replacing" originals or high-quality photographic copies get in the way of carrying out projects that will be useful for many researchers and, in fact, help extend the life of the originals.