Bell Labs is fading away, as perhaps all good things must. During its heyday it made countless contributions to the well-being of the world, largely by giving away knowledge. The enterprise was funded by an invisible tax on most of the telephones in the USA. A sliver of regulated monopoly profits were diverted to pay scientists, mathematicians, engineers, and a variety of other folks. They worked not on short-term moneymaking products, but on things that just might some day be important ... like transistors, information theory, and the cosmic microwave background radiation left behind by the birth of the universe. "Stuff", in other words, that doesn't show up on next quarter's profit-and-loss statement. Would that today's monopolists could make such contributions to humanity, and that dot-com dilettantes could lift their sights to such higher pursuits. Perhaps there's still hope for some of them.

But meanwhile, I owe a belated thanks to Bell Labs for a couple of small personal presents:

• Since time immemorial people have made tiles out of squares (or triangles, or hexagons, or other shapes) and have used them to cover surfaces. It's easy to make a repeating pattern, like a checkerboard ... but to make a nonperiodic tiling is quite another matter — a real mathematical challenge. Roger Penrose succeeded, in the early 1970's. I first read about Penrose Tiles in Martin Gardner's Scientific American "Mathematical Games" column (the source of a big chunk of my personal excitement about knowledge — thank you again, Martin!). I learned that Ronald Graham of Bell Labs had created sets of Penrose Tiles to give away, to help those who might want to experiment and learn about the phenomenon. So to Murray Hill, New Jersey I wrote, and some months later a package I received, containing a few hundred odd plastic shapes. I played with them and shared them with friends.
• Chess is complex. There are usually dozens of possible choices every move, and as the possibilities multiply together their exponential fan-out makes it impossible to see more than a handful of steps ahead. But near the end of a game, when there are fewer pieces on the board, things sometimes become more tractable. Kenneth Thompson of Bell Labs (yep, the same guy who created most of the original UNIX operating system) saw in the early 1980's that, with enough cleverness and computational power, a variety of interesting endgames could be solved — worked out entirely. So Thompson set to work doing "retrograde analysis", crawling backward from all the possible winning configurations that involved small numbers of chess pieces — King & Queen versus King & Rook, for instance. The result was a huge database that contained the absolutely best move in any given position. Ken put his results onto CD-ROMs and gave them away; he was kind enough to send me a set.

So, some decades later, thank you, Ron Graham, Ken Thompson, and everybody else at Bell Labs. I didn't come up with any breakthroughs from your gifts, but others did — and your generosity made that possible.

(see also CaissicMetaphors (8 Jan 2000), PersonalProgrammingHistory (2 Apr 2002), ThatDepends (9 Sep 2002), DeadBeginnings (28 Sep 2002), ...)

TopicScience - TopicPersonalHistory - TopicProfiles - 2003-05-26

(correlates: AnnotationPunctuation, ThanksAndAcknowledgements3, ClubScience, ...)