Storing and archiving design, service and production data is far trickier than it might initially appear. The original data often is computer generated, which means it must be stored in some electronic fashion. Consequently, a company may use everything from tape and diskettes to optical disks over time as convenient modes of filing. The problems with this storage situation arise when the data needs to be recovered later–often with some urgency.
One problem is that the media may have deteriorated, causing the files to become unreadable. This is common with serially archived bulk records such as magnetic tape. A small read error at the tape’s beginning may render the entire volume unreadable. The result: near total loss of data, a nightmare well known to those who have tried in vain to restore from a "backup tape."
Generally, magnetic media seems the least attractive for long-term storage. External forces can corrupt it, and magnetic media has its own self-decay mechanisms, which causes loss of readable information even when well stored.
Storage life of volatile magnetic media can be influenced by everything from ambient temperature to steel shelving. Also, the magnetic tape media exhibits serious mechanical degradation with repeated use, as well as "print-through" to the data physically underneath on a large tape with a thin substrate.
The same recovery problems exist with all diskettes. Many a critical file has been lost because diskettes were left on the dash of a car in the hot sun and became randomized or warped, or on top of a TV set or monitor to be casually destroyed by the cathode ray tube’s (CRT’s) degaussing coil.
Digital data has many recovery problems beyond media life. For one, the device or peripheral that can read the media may no longer work or be supported by current systems. This can happen much quicker than you would imagine. Five and a quarter-inch diskettes are an excellent example of media that was ubiquitous several years ago but virtually non-existent today.
Many cartridge tape formats and sizes of QIC, DAT or Travan type also have disappeared. Not to mention the essential demise of large-format reel-to-reel tape.
Optical media also had its share of false starts and orphan formats. Many early drives have vanished, from 128-Mb MagnetoOptical disks to Powerdrives. Optical media like writable CD-ROMs have hidden problems that can "bushwhack" the unwary.
First, the optical surface is written by laser heating, so it can be damaged by the same process (inadvertent drive operation) or careless storage (left on the dash of a car). The optical surface also can be scratched easily by normal handling.
Second, information may be written in a non-standard way that becomes unreadable as software formats and hardware change. The de facto standard CD drives for medium-size storage and transfer already are fragmented into many different styles and formats. As a result, the media often is not readable on other supposedly compatible drives.
You should also consider that new optical drives permit high-speed writing (2x, 4x, 8x), but the resulting image is simply not as well defined and reflective. And it often gives later-read problems on other drives. If CD-ROM platters hold important storage, they should be written at the 1x rate to insure the best platter image for later recovery.
DVD is yet another format destined to break into an unmanageable plethora of unhappy relatives. Each format offers some unique advantage but little compatibility with a standard–a bad outcome for us users.
Even assuming that the required peripheral still exists and miraculously works years later, the current application software and/or operating system may be unable to do anything useful with the stored data. This catches many companies by surprise. Old computer-aided design (CAD), word processing or design files may require a program or specific version that no longer exists in the system. Or the new version or "compatible" program may be able to read the old record but then does a horrible job of converting it. This is particularly true of PCB/Schematic tools, where library data can become different and incompatible over time.
Other taken-for-granted tools may quit working, as well, such as device programming software, which works only with a specific operating system. This can affect PROM/PAL programmers, equipment specific loading software, and various diagnostic tools. Old MS-DOS applications and drivers simply do not work in the new versions of Windows and can become a nightmare for downstream product support.
The rapid churning of operating system (O/S) software at Microsoft has been a pain in the neck for companies that now must deal with an unmanageable assortment of legacy files, drivers and programs for DOS (in all its myriad variations), Windows 3.1, Win95, Win98, NT3.51, NT4, WinCE, WinME and Windows 2000. And of course, all of those dependable "MS standards" are being abandoned in favor of a yet to be clearly defined Web-based O/S called Windows XP. If all that heartache doesn’t drive you into the arms of Linux, I don’t know what would.
Surprisingly, paper (at least the low-acid kind) has been the one data storage medium that has withstood the test of time. Printed copies of drawings, parts lists, procedures and manuals commonly remain fully functional long after the digital forms used to create them have failed. Thus, archives should always include printed copies as well as digitized documents.
Face it, you can read paper documents with no special tools. And paper has survived as a data storage medium for thousands of years and is surprisingly robust if well stored and protected. Hand illuminated Bibles from the Middle Ages and ancient Egyptian papyrus scrolls have fared far better than floppy disks and magnetic tapes–a sad commentary on the sometimes questionable "power of technology."
The digital past is a history of feuding proprietary technology and jockeying for market position. For the future, we need a durable computer media standard for long-term storage.
Walter Shawlee 2 may be reached by e-mail at firstname.lastname@example.org.