Project Names and Borges Numbers

Tom Van Vleck

This is an old story about project names. It was told to me by the late Arnold I. Dumey, a real old timer in the computer business, when he was visiting at MIT Project MAC in 1965. He published the first paper on hashing in the open literature, and held a key line printer patent. (Mr. Dumey was an Army captain in 1944, working at Arlington Hall for William F. Friedman, and was a member of the Voynich Manuscript Research Group.)

The story goes that a certain consultant visited IBM San Jose -- I guess this would have been in the early 60s -- to see somebody about a contract. This person was probably ex-IBM; at any rate he knew a lot of people at this IBM location. And on his way in to the office of the person he was to see, he met an old pal.

The friend said, "Oh, you're here on that Walnut thing?" The consultant said, "Mmm."

And when he ran into another friend, a few offices further, he asked, "What's up with Walnut anyway?" "Oh, we're into the acceptance tests."

And when he ran into another friend, he asked, "What's wrong with the Walnut acceptance tests?" "Oh, that thing will never work, the design's no good."

By the time the consultant reached his appointment, he knew all about Project Walnut. He knew how much it cost, how much over budget it was, and why it would never work.

Reporters do this sort of thing every day. It's neat, but not amazing. But when the consultant had finished his meeting, he said to himself, "Well, Walnut's a tree, it's something to eat ... and it's an exchange in the San Jose telephone directory." And he asked the first friend he encountered on his way out, "Say, what's the current status on BUtterfield?" And by the time he left the plant he knew all about Project Butterfield too, and how far over budget it was, and why it would never work either.

Now, I'm not saying this story is true.. but it's one of my favorites.

(I have often thought that it would be useful to create a list of names, chosen such that knowing one name on the list provided the least possible information about the rest of the list. We would, of course, call such an enumeration "Borges numbers," after the numbering scheme described in J. L. Borges' story "Funes the Memorious.")

(Walnut was an IBM information retrieval system built for the CIA and other agencies and delivered in late 1962. It used a Model 1630 photo-optical storage unit.)

Updated 02/05/95