On the Absence of Obsolescence

Five years ago, a colleague asked me where to find a history of computing that explored obsolescence. She believed, as many do, that computer technology had undergone the most dramatic progress of any technology in the second half of the 20th century, and thus also the most dramatic rate of obsolescence. Paul Allen, cofounder of Microsoft, put it this way:

It is possible that no other technology on earth has so continually renewed itself as computer technology. Advances in this field arrive in such swift succession that even the software and hardware of a few seasons ago are considered obsolete.¹

Surely, my colleague assumed, computing historians must have explored this issue. What were the exact conditions, she wondered, under which a computer was considered obsolete? How did inventors, designers, and users know when a technology was obsolete? Were such technologies still in use? Most importantly, how had obsolescence been used as a lens to understand technological change?

Unfortunately, far less has been written about this than she had hoped. Although a vast number of articles written for this journal mention obsolete hardware or software, typically the word is a token for “old” or “something better came along,” without any critical interpretation. My colleague and I remain unconvinced that these are the only useful or appropriate meanings.

Defining “obsolete”

At the root, obsolescence is a process that occurs when users chose to replace one technology with another, either because they need to (the original is irreparable) or want to (something better exists). That newer technology is also better is almost an aphorism of the last few centuries. Thanks to Moore’s Law, a newer computer is typically faster, smaller, and more powerful—for most people, this means it is better.

What makes something better can be difficult to pin down, however, as style, fads, and fashions can handily beat substance. Nearly 90 years ago Henry Ford discovered that in a stagnant market, the reliable, easy to repair, any-color-you-want-so-long-as-it’s-black Model T had difficulty competing against a more stylish and expensive-looking Chevrolet.² In the same way, the personal computer industry was a boring sea of beige until Apple introduced the iMac in 1998. The striking all-in-one design was round, colorful, and criticized for lacking a floppy disk drive, which many considered an essential peripheral. Nevertheless, it won over consumers and industrial designers, who flooded the consumer product market with candy-colored imitations. The plain, boxy PC has slowly disappeared.

Of course, what makes something better is often highly contextual. Depending on the point of view, an obsolete technology can be perfectly usable and acceptable for decades. Technological diffusion is never instantaneous, but an upward sloping curve of acceptance over time. Unfortunately, historians of technology have notoriously “skewed towards studies of the origins of technological change and not to its results,” and those behind the curve are often forgotten.³ Yet, when old replaces new, the now-obsolete technology provides the framework by which new and would-be users evaluate and interpret its replacement.⁴ The stories and perspectives of those who are otherwise “lagging behind” can unlock many mysteries for historians.

Consider the modern stored-program electronic computer, for example. Depending on your point of view, the world’s first such device came to life in 1949. Consequently, the entire class of electromechanical computers, including the six Ball Labs Relay Calculators and the Harvard Mark I, then became obsolete. This fate had been anticipated since at least 1946, when ENIAC was unveiled, but just how obsolete were these machines? All but the 1940 Bell Labs Complex Number Calculator were still operating 10 years later—after ENIAC was decommissioned in 1955, after the first transistorized computers in the mid-1950s, and after Fortran was already on its second version in 1958.

What had these obsolete devices been doing for a decade? Typically, they were executing highly repetitive calculations to generate mathematical tables, which were carefully published and then noted in the Mathematical Tables and Other Aids to Computation (MTAC), the clearing-house for such information since 1943. Throughout the 1940s and early...