A recent article by Nicholson Baker in The New Yorker ("A Reporter at Large," July 24, 2000, pp. 42-61) describes the trials and tribulations of the author as he tried to save old newspapers from destruction. Of interest to those who are concerned with archival preservation is the author's description of microfilming practices.
The author cited one particularly ironic example, given the prominence of microfilming in the world of government secrets: CIA documents that were so poor that the paper copies from microfilm were stamped "Best Copy Available" but were, in fact, "nearly indecipherable" (p. 48).
In the early days of microfilm, there were commercial interests that helped to drive the adoption of the technology, but promises about longevity were made without real certainty. For instance, the supposedly long-lasting cellulose acetate film used to support the emulsion beginning in the 1940s has turned out to degrade and has been replaced by polyester.
Today's polyester film is thought to be more permanent, but the test of time is much better than the tests of laboratories. Meanwhile, the emulsion that actually carries the image information (silver embedded in gelatin adhering to the film base) has its own problems. A 1991 Canadian study found emulsion damage in thirty-five percent of the sample microfilms examined (p. 56). Another study, in Florida, found that half of a particular collection had fungus problems (p. 56), and a librarian estimated that a third of one group of microfilms would need to be re-made (pp. 56-58).
Pointing out that fading of the original image is the biggest problem, the author cited his attempt to read a microfilm of a 1914 newspaper. "There were whole pages in which little more than the headlines was visible." (p. 58)
Even when images are readable, microfilmed newspaper volumes are considered "complete" if only a few issues per month are missing. Therefore, the presence of a microfilm copy of a newspaper is not a guarantee that everything is there.
Though Mr. Baker was lamenting the role played by microfilming in the destruction of original newspapers, there are important lessons for the larger archival community. Processes that purport to preserve must undergo serious tests - ongoing tests, not just tests that mimic the effects of age - and be subject to open, forthright discussion of benefits and risks, costs, potential ancillary problems, and so on. Those discussions must include the practical issues of how to check for copy quality, re-check for degradation over time, monitor proper procedures, and so on. Discussion must also be dispassionate and, as much as possible, untainted by commercial considerations. As the digital issues complicate archiving by bringing new and different variables into the equation, those involved in archival work must be open, forthright, and self-conscious in their evaluations of possible archival aids. To that end, we recently retrieved three CDs from our off-site bank vault storage location and compared the files on those CDs to the files on three originally identical CDs stored in the CSA office. This long-planned test showed that the CDs that had been in the bank vault (which is not air conditioned) for nearly three years were still in good condition. All the files were compared, and the files from the bank CDs were all identical to those on the CDs stored at the office. Monitoring of this kind is crucial to knowing whether we are accomplishing the archival work that is our aim.
When digital files are migrated from one data format to another, there should also be tests to demonstrate that the migration has been correctly performed, although those tests will be more difficult. CSA personnel have undertaken such tests when data have been migrated and have satisfied themselves that the migration was accurate. However, there are no simple tests if such issues. Scholarly expertise was required, as was a considerable investment of time. Nevertheless, the time was well spent. The preservation of digital files must be reliable - and be perceived as reliable - if we are to avoid the kinds of problems detailed in Mr. Baker's article.
For other Newsletter articles concerning the the use of electronic media in the humanities or digital archiving, consult the Subject index.
Next Article: AutoCAD® 2000 - A Review
Table of Contents for the Fall, 2000 issue of the CSA Newsletter (Vol. XIII, no. 2)
Table of Contents for all CSA Newsletter issues on the Web
Return to CSA Home Page