Add a Comment to this Note (list members only)
Dead medium: The Comparator; the Rapid Selector
From: boneill@allinux1.alliance.net (Bradley O'Neill)

Dear Bruce,

Here's some information on pre-encryption/decryption technologies of the 1930s and 40s. These creatures were the stillbirths of Vannevar Bush's projects at MIT and OP-20-G (Naval encryption division).

Most people know Bush as grandaddy of info-science, and prognosticator of hypertext (in the famous article in a 1945 edition of _Atlantic Monthly,_ Bush envisioned a hyper-linked bibliography system called MEMEX, an idealized machine that was never built).

Well, when I started looking into developmental background on BOMBE decryption devices for the German ENIGMA encryption system, I stumbled onto a source examining Vannevar Bush's role in creating Rapid Selector/Tabulating machines for the Navy and private industry, all inventions that predate Bush's idea of MEMEX.

This particular text is I'm citing is _Information and Secrecy: Vannevar Bush, Ultra, and the Other Memex_ by Colin Burke; Scarecrow Press, Inc. Metuchen N.J., 1994. LOC: HD9696.C772B87 1994

Dr. Burke goes in-depth on several Bush "Rapid Selector" inventions that precede the development of successful analog optic-cryptoanalytic machines of WWII. Principal among them:

THE COMPARATOR: 70mm Eastman-Kodak paper-tape based electronic crypto-analytic prototype, funded by the US Navy, built mostly at MIT, first assembled in 1938. The Comparator was plagued by years of mechanical setbacks. Bush wanted a "high-speed" (projected to be 100 times faster than 1920s tabulators) parallel processing analyser that utilized photo-cell light readings to index (and thus decode) up to 50,000 character comparisons per minute.

Very low memory capability caused printing/retrieval problems. Bush realized that without microfilm density, the processing speeds were also unachievable. And if microfilm was used, then the reading/recording capabilities would suffer from insufficient resolution.

THE RAPID SELECTOR: Begun in 1937. Bush's MIT team first built this analyser in 1940. Funding was dropped by a disgruntled FBI and subsequently picked up by various private foundations including Eastman and NCR (Bush was apparently an undaunted spinner of techno-dreams ala Steve Jobs). The Rapid Selector went through several incarnations, but was conceived as a specialized data- retrieval system for business records or scientific research.

The Rapid Selector was a microfilm-based analyser consisting of a 7' tall relay rack, housing the film drives. Like its sister,the Comparator, it used a light- sensing reader system to allow speedy retrieval of microfilmed information. The user compiled a series of punchcard notes that were indexed into microfilm storage by a system operator/librarian.

The Rapid Selector would then allow the user to cross-reference other researchers' additions to the user's "specialized area" without sorting through irrelevant texts. Bush saw the Rapid Selector as an eventual replacement for card catalogues.

Although Bush conquered his basic speed/retrieval problems, the required coding system to access information ultimately proved prohibitively complex. The specialized typewriter for the code-punch was also unworkable.

Burke's text is full of other useful information, follies, and successes that orbit around the development of these pre-digital machines. I'll post more as I digest it.

Regards,

Bradley.