A Partial History of Computer Conferencing in Ann Arbor

originally posted: June 26, 2007
last updated: October 23, 2007

I originally wrote this article in 1997, with occasional updates in the ensuing years. I think the "Great Green Room" is a better home for it, so I'm importing it here.

Introduction

That's the whole problem with trying to document the history of computer conferencing - it predated the Internet, so it developed separately in different places. So there isn't one history, but many histories. Computer conferencing had many different beginnings, that have only in the last few years been converging.

This article talks about one of those beginnings - the one that took place in the city of Ann Arbor, Michigan. Starting in the research labs of the University of Michigan, and moving out into the surrounding community, Ann Arbor's conferencing systems were among the first to make sophisticated computer conferencing systems publicly available. The software developed in Ann Arbor, and many of the ideas incorporated in it, have been extremely influential and have been much copied. At the same time, Ann Arbor's systems have a history of dedication to free public access and to democratic control that remains unique world-wide.

Howard Rheingold has called computer conferencing systems like those discussed here ``Virtual Communities.'' The first computer conferencing systems were designed with the idea that they would be used primarily for teaching or for professional or business discussions. But as groups of people started talking over them, they began to form genuine on-line communities - people sharing their lives with other people they had often never met. It was discovered to be something that many people considered a pleasure in its own right - a way to talk about any topic, deep or shallow, with people willing to listen and respond. A way to meet people with diverse viewpoints from all walks of life. So this document is not only a history of some computer systems, but a history of some communities.

CRLT and Confer

Confer pioneered the structure that has become the standard on most modern computer conferencing systems. Any user could introduce a discussion topic (called an item), and other users could append their own responses onto the end of it. Responses to each particular item are kept in order, and the system remembers which responses each user has seen, so you could request to see all items and responses that you haven't seen before. There were separate conferencing areas for separate topics, but very few (you had to run a different copy of the program to read a different conference). Confer's direct descendents, Marcus Watts's PicoSpan system and Charles Roth's Caucus system, were more flexible about offering large numbers of different conferences.

By the time I started using Confer, around 1982, there were several very active conferences on MTS. Discussions were predominantly computer related. The USER:FORUM conference provided a forum for questions and answers about using the MTS system, and the CRLT:MICROS conference was organized by Karl Zinn to provide an area to talk about the latest microcomputers then appearing (I remember rather wild rumors about what IBM's first microcomputer was going to be like). It was a wonderful way to keep up with the rapid changes in the technology. But over the years people discovered that computers could also be used to talk about things other than computers, and more and more general social discussion arose. By the late 1980s, the shift to more social conversation was marked by the appearance of MEET:STUDENTS.

Karl Zinn continued to serve as an advocate of computer conferencing and other forms of computer-aided education. He provided encouragement and a home at CRLT for a number of other people interested in developing computer conferencing systems. Robert Parnes and Charles Roth started a project to port Confer from the MTS operating system to Unix. The project was never completed, but Charles Roth was sufficiently inspired to develop his own Unix-based conferencing system called Caucus, which he developed into a product that he marketed for many years. Robert Parnes eventually went on to write his own Unix version of Confer, Confer U, which is also still in use today. Also at CRLT, Marcus Watts and Joshua Fein developed a system called MicroSpan.

The Early Days of M-Net

Mike had been running a small home bulletin board on a Atari 800 with one dial-in line. His system was typical of the many small bulletin boards that were beginning to sprout up around America at the time. But Marcus Watts soon sold Mike on the idea of setting up a much bigger system, which would be open to the public like most small bulletin boards, but which would have software similar to Confer, and would be more ethically administered than SEMCO's system had been.

Mike purchased an Altos 68000 computer, a single-board computer that ran System III Unix and that had 16 serial ports built in and could thus handle a large number of dial-in lines. Mike paid about $30,000 for the machine, and much more for various upgrades and spare computers over the years. By any standards, spending that kind of money on a computer, and then just opening it up for the public to use was an amazing thing to do. Generally speaking, machines that powerful were available only to people at universities and businesses, not to the general public.

Marcus set to work writing software for the Altos. This included three major programs:

• The newuser program, which would automatically create accounts for people who logged in. This was inspired partly by many BBS systems of the time, partly by a program on a system on the Arpanet at MIT which would automatically create "free" accounts, and partly by a program called "newuser" which Marcus saw running on Fortune 32:16's at tradeshows.
• A party live-chat program which was inspired by the "CB" program on Compuserve that Marcus saw briefly.
• PicoSpan, a Confer-like conferencing system, which was originally meant to be light-weight version of MicroSpan.

M-Net represented a unique combination of two previously existing movements. Half of its roots were in the university research community, where sophisticated computers and software were being used to study electronic communications behind largely closed doors. The other half of its roots were in the vibrant BBS and computer hobbiest community, where teenage kids ran small microcomputer-based message systems on their home computers, allowing random strangers to dial in, and never dreaming of charging anyone money to do so. M-Net combined an unusual amount of technical sophistication, with an playful, wide-open approach.

As a hobbyist system, M-Net didn't validate users, or charge any fees. At first it simply didn't occur to anyone that it would be possible to charge money. Mike paid all expenses out of pocket. M-Net very quickly got to be very popular, drawing users from the local universities, the high schools, and the community at large. And it very quickly became a whole lot of fun. Mike ran the system as a very benevolent dictator, describing his role as ``M-Net Janitor'' and leaving users free to do very much what they liked.

By this time I had joined M-Net's staff. Around 1984, I started donating money to M-Net to allow it to grow faster than Mike Myers's generousity alone would allow. We formalized this idea, allowing users to become patrons by donating money, and giving patrons access to a few additional phone lines beyond those accessible to non-paying users. Russ Cage became M-Net's first offical patron. M-Net was able to become a break-even operation, even slowly repaying a bit of Mike's original investment.

Originally, M-Net was accessible only through its gradually growing pool of local Ann Arbor dial-in modems. But M-Net was able to gain users outside Ann Arbor earilier than one would expect by piggy-backing on the the University of Michigan's statewide computer network, the forerunner of today's Merit network. At that time, Merit maintained pools of publicly accessible dial-ins in many Michigan cities, and some publically accessible dial-out lines in Ann Arbor. M-Net soon started getting users from many parts of the state coming in over the network, with large communities of users in Houghton and Lansing. (The friendships these remote users made in Ann Arbor through M-Net eventually lured several of them to move to Ann Arbor.) After the University grew unhappy with the large number of calls their dial-out modems were making to M-Net, Jon Zeeff established a more formal direct connection from the Merit network to M-Net (though this was limited to three simultaneous users).

M-Net also used a program called "remps" (written by Jon Zeeff) to share PicoSpan conferences with several other systems, including FishNet, an M-Net like system that Mike Myers had started in Florida, and Chinet, a system in Chicago run by Randy Seuss that probably predated M-Net, but which by this point was also running Picospan.

For a short period before the founding of Byte Magazine's BIX system, M-Net hosted a "Byte Magazine Conference" in which several of the editors and authors participated. (Jerry Pournelle entered several items with titles like "Bee-dee bee-dee bee-dee" complaining about the incomprehensibility of the PicoSpan interface.)

M-Net soon grew to be the center of a large amount of face-to-face social activity, as well as on-line conversations. Quite early in its history M-Net started holding monthly ``PicoFests,'' face-to-face gatherings at local restaurants, followed by ``Post-PicoFest Parties'' hosted at the homes of users. In the spring of 1986 Denise Anderson launched a series of weekly M-Net happy hours on Friday nights, held at local bars (these were always unpopular with the bars, because large numbers of people would show up and just talk, without ordering much food or drink). Around 1989, users began to meet each Saturday morning for a walk in a local park. Various spin-off groups, such as the M-Net Bicycle Touring Society and the M-Net Volleyball League made appearances. Face-to-face events always seem an important part of on-line communities.

The Early Days of Arbornet

Sometime in 1984, Marcus's employer, a Ann Arbor company called Network Technologies International (NETI), had purchased the rights for PicoSpan from him, planning to develop it into a commercial product to be called E-Forum. NETI was an odd company, founded on a combination of 60's idealism and the dream of making a quick buck. One of NETI's projects was to try competing with nation-wide on-line services like Compuserve by starting a large number of small, regional community conferencing systems, which would be similar to M-Net, but would be commercial for-pay operations. The first test of this concept was launched in late 1984(?) in Ann Arbor, initially under the directorship of Jeff Williams. It was called Arbornet.

Arbornet was supposed to be a profit-making enterprise which people would pay to use. This was a somewhat surprising and slightly repugnant notion to many of us using M-Net. Half the fun of M-Net was being free. Unsurprisingly, Arbornet was not a commercial success. Being in the same town with M-Net caused problems for Arbornet. It is hard to make money selling something that other people are giving away. Arbornet remained small and quiet, while M-Net continued to grow.

In April of 1985, NETI founder Larry Brilliant and the publishers of the Whole Earth Catalog started another regional for-pay PicoSpan system, this time basing it in San Francisco. They called it The Well. The Well has been lively and influential and is still very much alive, but it was not a significant money-maker for NETI. Many of the NETI employees who helped create the Well had been long time users of M-Net and Arbornet, but there was almost no continuing interaction between the Well and the Ann Arbor virtual communities (aside from a brief period when Mark Ethan Smith found a home on M-Net after being ejected from the Well). The cultures of the two systems are very different.

Around the same time, NETI decided to abandon Arbornet as hopeless. It wasn't making money, and didn't ever seem likely to make money. However, Arbornet had attracted a small group of loyal users. Rather than see Arbornet vanish they formed a non-profit corporation to take over its operation from NETI. The key inspiration behind this probably came from Susan Holtzer (now a mystery novelist living in San Francisco, but then a politically active Ann Arbor resident who had gotten sucked into computer conferencing on M-Net by her son Mason Jones). She envisioned Arbornet as an on-line, grass-roots community communications center. Among the people that got interested and involved were local lawyers Don Koster and Dave Cahill, Carolyn Clock Allen (an early Confer user and a NETI employee), and U of M linguistics professor John Lawler. NETI provided substantial support for the fledgling non-profit Arbornet, including funding and hardware.

Arbornet was almost certainly the world's first member-owned and operated virtual community. It remained much smaller and much less active than M-Net though. A year or two later, Arbornet obtained 501(c)3 status, making it a tax-exempt organization. It remained a for-pay system but, in practice, it didn't collect much money.

Arbornet, in fact, outlasted NETI. NETI had been involved in some other moderately successful conferencing projects (including Byte's BIX system and GE's Genie), but on the whole, mostly lost money. NETI's owners apparently at some point decided that there was more money to be made in stock market manipulations than in computer conferencing, and the company eventually disappeared in a flurry of law suits. There was apparantly some disagreement between Arbornet and NETI about whether or not NETI had fulfilled it's promises to Arbornet. Fortune Magazine posthumously awarded NETI the distinction of having been ``the biggest stock market scam in the the history of the Vancouver exchange.''

The Selling of M-Net and the Founding of Grex

Dave's term as M-Net's owner was rather troubled. He had paid a lot of money for it, but the sudden sale had undermined the confidence of the users, and donations were slow coming in. He was drawn into endless fights with users who didn't agree with the way he was running it. Several times he threatened to shut down the system if more donations didn't come in. Users proved unwilling to donate money to a system that might be shut down at any point. In January of 1991 Dave posted an article saying (1) the guest lines will be shut down, (2) newuser will be shut down and M-Net will be strictly a for-pay system, and (3) he had just posted an article to USENET advertising M-Net for sale.

This message galvinized the long-standing discontent of some of the users. Four days later, a group of long time users (Marcus Watts, John Remmers, Mary Valdivia, Steve Andre, Misti Anslin, Tom Doehne, Valerie Mates, Marc Unangst, Denise Anderson, Fred Sleator, Brian Dunkle, and Mike Smerza) got together to plan the launch of a new open-access system that would be user-owned instead of privately owned. This system would eventally be called Grex (the Latin word for "group").

Dave backed down on all three of his announcements (largely because of Meg Geddes' efforts at negotiating a peace) and the Grex group postponed their plans, hoping M-Net's problems could be resolved. But problems continued. At one point Dave deleted the accounts of several users, including Marcus. The Grex founders decided to go ahead with their plans.

Grex came on line on June 26, 1991 and opened to the public on July 18. It ran on a Sun 2 computer contibuted by Mike Bernson. It had four public dial-in lines, and was housed rent-free in a warehouse belonging to Ken Ascher. A non-profit corporation called Cyberspace Communications had been formed to run it. Grex was firmly dedicated to open access, pushing the idea further than M-Net had ever taken it. They abandoned the idea of offering extra dial-in lines to paying members, not wanting privileged classes of users on the system, and they lowered the membership fee to less than half of M-Net's. To ensure democratic operation of the system, they wrote bylaws allowing any member to call for a binding referendum on any issue.

The M-Net/Arbornet Merger

Iain O'Cain was at that time also involved with Arbornet, along with its president, Jeff Spindler, and Marae Price and Dave Lewis. Arbornet still had some money, but had almost no users. But they did have that attractive 501(c)3 tax exempt status, plus bylaws and five years experience running a member-based system. So Iain suggested a that OAFS and Arbornet merge, combining M-Net's working computer and larger user base with Arbornet's working organization. This occured in 1992 (or late 1991?), and the Arbornet non-profit corporation took over the operation of the M-Net computer system.

Since Then

Former M-Net users and new users either shifted to Grex or stayed on M-Net as they pleased. The Friday night happy hours remained an M-Net event, but the Saturday morning walks evolved into an Grex event. Though the systems continued to offer similar services, their online cultures gradually developed quite distinct personalities.

M-Net continued to grow under its new administration. It finally replaced its now ancient Altos 68000 with a more modern 486 system running BSDI Unix. It also replaced the Picospan conferencing system with a clone written by Dave Thaler called Yapp.

M-Net rented office space in Ann Arbor's NEW Center, which is an organization that encourages the growth of non-profit corporations, and began providing some computer services to other NEW Center tenants. (Grex moved into M-Net's former location, a room called "the dungeon" in the basement of the 110-year-old house Dan Byrne had owned when OAFS bought M-Net from Dave Parks.) In December of 1992, M-Net obtained a full Internet connection. Arbornet won a grant to create a conferencing system for K-12 teachers to be called TeacherNet.

However, M-Net continued to be troubled by a high level of conflict, mostly over what its mission as an educational and community service organization ought to be. TeacherNet starved for lack of volunteers willing to work on it. Most of the users of M-Net had little interest in community service. They just wanted to use the conferencing system. Other users wanted to see Arbornet pursue a loftier mission. This fundamental difference in goals led to years of conflict, and many resignations from the Arbornet board of directors. It has still not been fully resolved.

Generally M-Net's culture has become rather rough and rude. It's a strong community, but not one where anyone with a thin skin will last long.

For many years Arbornet was on a reasonably sound financial footing, building up decent reserves of money. However, around 1996 there was a period of time when financial information stopped being reported to the users. Later when the the books were revealed, it was discovered that Arbornet was in a financial nose-dive. With high costs and little income, it was losing money fast. Doom was predicted, and continued to be predicted for years, as a small number of users, continued to donate just enough money to keep paying the bills. Under the leadership of Dave Cahill and with the support of Todd Plesco, Joe Saul and others, Arbornet has been able to reestablish solvency. To cut costs they eliminated their office and moved their machine into space owned by an World Wide Net, an ISP based in Livonia Michigan. Only four dialins remain, but they are local to all of south-east Michigan.

Grex has continued to thrive, upgrading first to a Sun 3, and then to a Sun 4/260 and a 4/670. It open up an internet connection early in 1993 and gradually more and more of its users came in over the internet. By 1999, it had over 25,000 users, with more than 200 new accounts being taken out every day. After some disputes with the city of Ann Arbor over the height of the ceilings in "the dungeon," Grex moved to a room in an old industrial building. This room was dubbed "the pumpkin" after its curious wall coloring. Most recently, in December 2004, Grex was moved onto a x86 box running OpenBSD and the machine is now located at an ISP in Ypsilanti, Michigan.

Grex pushes the open systems ideology a bit further than M-Net, and though it has never been as wealthy as M-Net was at its height, it has never been as near financial ruin as M-Net was more recently. Discussions there tend to be much less adversarial (it is considered by some M-Netters to be rather boring). The Cyberspace Communications board works very much by consensus. Big decisions are always made by user referenda. The board usually won't act on even small decisions until they are discussed to death in a public, on-line conference.

Though the two systems started out virtually identical, they have diverged nicely, with two quite distinctive cultures developing on them.

Both systems still provide free Unix shells. These days however many people use these system solely for E-mail. Giving free E-mail is a worthy goal in itself, but it not what the systems were created for, and having tens of thousands of E-mail users is a significant burden.

Both systems have developed web interfaces to their conferencing systems, while retaining their old command-line tools. Grex runs Backtalk for the web interface and Picospan for the command line interface. M-Net used a web version of Yapp for many years, but has replaced it with Backtalk. It still uses Yapp for the command line interface.

M-Net and Grex have inspired a large number of other systems in the Ann Arbor area. Meg Geddes' Netmeg system started as a BBS in 1985 and evolved into a free provider of Email and newsfeeds for local systems, including, at times Grex and M-Net. Bruce Wagner ran a system called MetroPlex for some years, and Dave Simmons ran one named InaNet. A PicoSpan system called AIX (Annarbor Information eXchange) focused mostly on CPM tips. More recently, Jared Mauch started the Nether.Net system, which gives free shell access to all comers. The Huron Valley Community Network was originally inspired more by the freenet model than by Grex or M-Net, but there has been a long history of close cooperation between HVCN and Grex.

In 1999 both M-Net and Grex took part in an ACLU lawsuit, Cyberspace vs Engler, aimed at overturning a Michigan law that would severely penalize anyone making sexually explicit material available to minors over the internet. Such a law would make it impossible to offer uncensored forums for free speech on the net.

All of the Ann Arbor systems differ strongly from others I have seen. First, they have more younger people. That's a natural consequence of being free. This can be trying, but it can also be a great pleasure. We've seen a lot of people grow up on our systems. We've seen shy introverted kids acquire social lives, we've seen people who could hardly type a sentence become articulate writers, we've seen pesky trouble-makers who kept trying to crack system security become valued staff members. That's nice. Having some kids to raise makes you more of a community.

The Ann Arbor systems also see a higher incidence of inane chatter than most for-pay systems. It'd be nice to blame that on the kids, but many of the biggest babies on the system are in their thirties or forties. I think to some extent people take things they don't have to pay for a bit less seriously. Over all, open access works far better than one would expect.

There have been other free-access conferencing systems now (for example, Utne Cafe and the mostly deceased Electric Minds). There is even another user-governed system (The River). But the only systems that are both free-access and user-governed are still the ones here in Ann Arbor - M-Net and Grex.

Acknowledgements