Robin Hanson's "Idea Futures" is a collection of ideas in the WWW, a betting market on important questions in science and technology.

"Idea Futures" (http://skyler.arc.ab.ca/~jamesm/IF/IF.shtml) began as a glimmer in the eye of Robin Hanson, but it took SeanMorgan, Mark James, and a diverse group of "techies" (scientists, engineers, programmers, etc.) in Calgary to make it the reality it has become.

Act I

In 1984, Robin Hanson left graduate study in physics and philosophy of science atChicago to associate with a community of idealistic techies who shared an exciting, if unconventional, vision of a high-tech future filled with nanotechnology, artificial intelligence, etc.

But what most attracted Robin was their hopes to revolutionize how academia, andsociety, evolves ideas ­ by developing a hypertext publishing standard wherein it would be easy to find all criticism of any archived work, and which supported personal reading filters built from decentralized reader evaluations. (The WorldWide Web is now the main hypertext publishing standard, but has none of these features.)

In 1988, however, just as this project was starting to be funded (as Xanadu, Inc.), Robin developed doubts about how well considered this vision was, and sought alternatives. He soon imagined "Idea Futures", essentially betting markets on important science and technology questions. Robin and friends expected that those who were willing to "put their money where their mouth is" would back their group's views against knee-jerk academic critics and other clueless pundits.

Robin also soon realized that subsidies of such betting markets could induce research toward answering science questions, even when such markets would otherwise be very thin. Such "information prizes" could make researchers more accountable to their research patrons.

Robin spent the next four years of his spare time researching, writing several papers, and giving many talks on this idea. (He continued his day job in artificial intelligence research.) Robin, however, didn't have the connections or money required to overcome the gambling laws which still outlaw his dream. So instead, he developed a related board game, which he failed to market, and worked on an email-based play-money "Idea Features" game, where he spent too much time developing fancy internal features, neglecting the interface and just getting it going.

About 1992, Robin burned out on the topic, and turned to inventing alternative ways to run conversations, to buy health care and law enforcement, and more. But, without credentials or contacts, and with two children filling up his spare time, Robin despaired of getting anyone with power to listen to him. So in 1993, Robin returned to graduate school, this time in social science at Caltech, hoping to become a credentialed expert in institution design.

Act II

In 1994, David McFadzean brought a 1992 "Extropy" article of Robin's to the attention of the BioSimulation (or BS) group, a weekly informal discussion group meeting at the grad club bar at the University of Calgary. This group consists of David, Mark James, Mark Hammel, Ken Kittlitz (all once computer science graduate students), Duane Hewitt (a genetics grad student), and Sean Morgan (a manufacturing engineer with a degree in engineering physics).

Originally formed to discuss topics such as genetic algorithms, complex systems, artificial life, and neural networks, recent topics included nanotechnology, quantum computation, the singularity, and memetics.The BioSimulation group shared many links to Robin's community of idealistic techies (though Robin has yet to meet the rest of them in person). David and Ken, like Robin, had long subscribed to the journal "Extropy", and were members of the email list "Extropians@ extropy.org", where Sean also lurked. And Sean had previously helped develop the Nanotech Web pages (at"http://www.arc.ab.ca/@morgan/ Nano.shtml").

David showed Robin's "Extropy" paper to the BS group with the idea that they might play a paper-based version of it. But the group soon envisioned a game implemented on the World Wide Web, and decided to try and make it real. On July 31, Sean emailed a delighted Robin that they had "started working on a WWW server for [an] Idea Futures" play-money game. Their two main reasons for doing this were a fascination with "Idea Futures", and technorapture regarding the Web.

In reading about "Idea Futures", they were impressed with its ability to cut through the noise surrounding debates on scientific and technological issues and come up with an honest consensus on their likelihood. At the same time, the group could test their abilities to predict trends.

By implementing this on the Web, they could learn more about these new and exciting Internet protocols. A Web server was more appealing than an email interface, because it provided greater interactivity. And having just been exposed to Web forms for the first time, Sean underestimated the effort required by at least an order of magnitude.The development was done by Sean Morgan and Mark James at the Alberta Research Council (ARC), their employer. Robin gave them what help he could from long-distance, sending on August 9 a summary of his prior email game design, and the actual LISP code, and offering opinions by email on many issues duringdevelopment.

Sean and Mark's roles could be characterized as "Forger" and "Honer", to Robin's"Inventor" role ­ and all three roles were critical. Sean was usually interested in adding new features to the system, whereas Mark has worked on system performance issues such as access security, maintenance of backups, efficiency and reliability.Sean developed the system design and user interface, and wrote a code that might be functional, but crude; Mark improved its efficiency and tracked down the more subtle bugs. Before each new major task, Mark would help Sean to find and understand an example in an on-line library. Then Sean would beat it into shape ­tying it into the existing programs and data files and providing a user interface. Finally, Sean would pester Mark to fix the problems Sean's modifications had introduced.

A fourth critical role was played by the many "Idea Futures" mailing lists, which were created and maintained by David McFadzean. On these lists, members honed thewording of claims prior to their creation, elaborated them during play, and provided high quality discussion of related issues. These mailing lists also served as a forum for discussing enhancements to the game.

Sean and Mark made surprisingly rapid progress. Idea Futures was by no means an official project of the ARC. And before they could make progress, Sean and Mark had to learn Perl (the development environment), http daemon configuration, advanced html and forms inputs, and how "Idea Futures" was really supposed to work.

Yet by August 15, Sean and Mark had an alpha-test version of the game ready. And on September 23, a beta-test version was announced to the Extropians mailing list. Within a month there were about 30 users, and since then the number of users had doubled about once a month (at least until today, with over 600 members on April 11 1995). Players could add their own questions to bet on by October 6, and automatic postings of the going odds to newsgroups began on November 25. On January 20, 1995, Sean left the ARC to join SAP-AG, and Mark took over the system administration.

Act III

The ARC has recently become interested in the system as a tool for technology scanning. As one department head says, "it's a cheap way to get the opinions of 600 people". So they will soon officially allocate some developer time to enhancement of the system.

The "Idea Futures" game has a long "To Do" list. But the first priority is clearly to find a way to handle continued rapid growth. Work is under way to redesign the server to handle more users on a single server, and to allow the market to be distributed across many machines.

And looking further into the future, everyone hopes that this game will inspire someone somewhere to try "Idea Futures" with (real) money.