In 1990 a researcher at Apple named Harry Chesley (who later went on to do Shockwave and now works at a software firm
in Redmond, Washington) ran across a technical paper from Xerox PARC® in which other researchers had published an al-
gorithm for updating distributed databases. Chesley used the algorithm and crafted a small application for Appletalk net-
works. He dubbed it Rumor Monger and gave a few copies to friends in the lab. After several iterations, he released a
rather stable and powerful free version that would not time out after X days or weeks. It worked and it would not stop
working until the networks were shut down.

HERE'S WHAT IT DID:

Every Macintosh include a local area network, and you can easily hook two Macs together, and they will begin network-
ing. What Chesley called Rumor Monger at work was known as Lightweight Asynchronous Conferencing System (LACS)in
his technical article, partly because of the turmoil it caused in the company when the head of the Advance Technology
Group, (who is now chief technology officer at ATT) got involved.

LACS or Rumor Monger runs in the background. It has a small window where messages of 255 characters or less can be
composed, with an expiration date and then propagated around the network—-anonymously. If you were looking for a ride
home one evening, you would not keep sending the message for more than a few hours. The user had two other windows
showing messages unread and those already seen. The algorithm tried to propagate hot rumors (new ones) and slowed
down the spread of old rumors. The status window showed what machines were communicating, but this feature was added
later, after the controversy (as | recall)over anonymity. Up until then it was anonymous, and messages could not be traced.
The Xerox algorithm allowed the program, the system to keep tabs on all these rumors floating around the Apple research
network.

Chesley's article goes into great technical detail, but he admitted the more interesting part was the way it was used, the
reaction by the executives, and the surges and low traffic times for the program. Most people gave others a copy and never
removed it unless they did a complete system upgrade or installation. At the early stages, everyone knew the people who
had the program, and you could even guess the writing style of some mongers and try to imitate them-all in 256 charac-
ters. | remember standing outside a friend's office and laughing and reading rumors together. As the funny ones propagat-
ed, ripples of laughter would bounce around the large office as the rumor self-propagated. After a while, the use of it was
not as amusing or interesting, but any time there was a pending coup d'etat in the executive ranks, or some company was
trying to buy Apple, everyone turned to Rumor Monger to find out what the P.R. department (whose newsletter was known
as “Pravda") would not tell us --if they even knew! During any period of anxiety or change, the rank and file would turn to
Rumor Monger for comfort and amusement. It was incredibly useful, comforting, and full of half-truths. Unfortunately, only
a few could tell which half was true.

The 170 kb program quickly spread around the company, and when it reached the sales and marketing arm, the users
showed less restraint in the way it was used: ad hominem attacks, mean spirited lies, and even libelous rumors were
spread about co-workers and upper level managers. In spite of the code being published by Chesley and the program
available by ftp form ftp.apple.com (it's not there now), we never heard of any other Mac campus or business unleashing
rumor monger on the social structures of a university or Fortune 500 company. Apple's social structure was relatively
loose compared to some of the authoritarian regimes that were Apple's customers. What would have happened there?
However, the code awaits the next decade of programmers who may wish to craft a tool for the rumor mongers of the 21st
century to use and to see its effects on furthering Post Governmental Disorganization.

The full technical details and source code are in Harry Chesley's article in develop_ Issue 2 1991. develop_ was a
prin/CD-ROM technical publication that is no longer published.

..and the original algorithm: Alan Demers, Mark Gealy, Dan Greene, Carl Hauser, Wes Irish, John Larson, Sue Manning,
Scott Shenker, Howard Sturgis, Dan Swinehart, Doug Terry, Don Woods: "Epidemic Algorithms for Replicated Database
Maintenance", Xerox PARC Technical Report CSL-89-1, January 1989.

— eee...