Optical disks: Are archivists repeating the mistakes of the past?

Abstract

Practical, workable, and relatively inexpensive optical wstems have piqued the interest of ~me archivists In the capabiIitles of this new technolo~.l. Promises of rapid ~arch and retrieval capabilities with cheap storage costs and a stable medium seem especially tempting to repositories with large ~IIections of graphic images, maps, drawings, and textual ck~cuments. As archivlsts debate using optical .~/stems for storage and retrieval of records already in archives, we must not avoid the question of how to Identify, acc~ess... and preserve information created elsewhere and stored m optical based systems. Archivlsts have rarely stopped to consider what we wi II do with information from optical disk systems that are created outside archives by government age.holes, corporations, institutions, and mdlvlduals. But we should r~.c)gnlze that the same features which make optical disk te~hnolog/,, appealing for use in archives ( high density storage, rapid access, imaoe enhancement, easy reproduction, etc. ) also make this technology., attractive and cosleffective for many applications in oraanizations with ~tive records systems The records that offices are ~annlng onto optlcal disks t_ndav will confront archivists .~oner or later as another information technol .ogy problem unless we beoln developing strategies now to preserve information in this format. No standards, guidelines, or procedures exist to held archivists handle records stored on optical disk .~/stems, but there are important similarities between optical and m~netic e/stems. Archlv)sts can apply some lessons learned from our previous exDerlence with magnetic medla systems to respond more effectiveIv to the Introduction of optical disks. The first los, son to learn from our experience with maanetm electronic records is that the ;mort life span of storage media, although an important problem, is not the most challenging asp~t of preserving records of modern information s./stems. Archivists who perceive me threeto ten-year useful life o[ magnetic media as the main problem of electronic. records are likely to consider the lon~r proje~cted life of optical dlsF:s as a s]~niflc~nt advance or even as a posslble solutlon to many of the problems that archivists confront ',vhen pr~ervmgcomputer-generated records The optical medium m~¢ last 30 years or lancer, thus reducing the frequenc¢ at wnlcn tne mformatlon must be c~pled onto a new storo~ device, and optical disks cannot be erased m the .:arise that magrletlc media can. To be sanguine about optically stored mformatlon as a consequence, however, over ]oaks the much more crucial Issue of system dependency which opticaland magnetm systems share. Optical ~stems pose many of the same access and preservation problerns,~s m~netIo systems and they rai~ some new issues of t.nelr own. Because both types of systems rely on software to retrleve information, archivists must preserve both the medium on which the records are stored and the means to access me records ( hardware & software). Even if, indeed especially because, optical media may last for fifly years, there is no assurance that the hardware and software necessary to retrieve rec~)rds from that medium will be usable or available in fifty years. Anyone who has encountered wire recordings, wa× cylinders, punched paper tape or Mc~,ee cards, to name only a few obsolete information storage technologms, can apprecmte this problem on a small scale. Archivists who work with magnetic electron;c records try to mlnlmlze the need for software by reformatting electromc mformatlon into a softwareindependent format. In the ab~nce ofwldely adopted data and document Interchange standards, this approach has been effective for certain types of machine-readable data files. '~et some archivists are becoming uncomfortable with this practice. The strategy, of creating software-lndependent data for arcnlval preservation is increasingly hmlted because information from most document-based and m~a~le-based systems cannot be retrieved or interpreted In a .~ftware-Independent format. Thi.~ problem is even more apparent with optlcal systems because there Is no such thing as a ~ftwareindependent for m at for