Published in the May 1993 issue of Dr. Dobbs' Journal
by Lee Felsenstein
Interval Research Corp.
One of the most important factors in the development of personal computers has been the urge to permit people to connect, for whatever purpose may suit them. Where does this come from, and where is it going? How important is it? What does it mean in thi nking about society in cyberspace?
I will argue here that there is a "commons of information" which has been fundamentally important in the survival and development of mankind, that it has been minimized in our present-day society, but that the inbred need for the functions provided by suc h a commons are very much with us. I will further argue that the development of personal computers has been seriously affected by the quest for the revitalized commons of information, and that this quest is showing signs of success.
I refer to the commons of information by the Greek word for the city-state's place of assembly: the agora.
It is not natural for people to live in isolation from each other. All traditional societies are based upon the village or at least the nomadic clan. All villages are centered around some space of assembly, which usually functions as a marketplace.
What goes on in these marketplaces is more than commerce. People hang out there, display their identities (usually as members of groups), gather groups of friends, banter and gossip within and among the groups, overhear others' conversations and inject t hemselves temporarily into those conversations. In short, they get to know who the other people are who share their society, and keep up with their daily doings.
Does this sound a bit like life at the mall? If so, there is good reason. The mall, like the village square, is a space where transactions of various sorts are carried out in public. These include commercial transactions, social transactions, and politic al transactions. Obviously, not the entire transaction is carried out in public, but they are set up so that they can be completed in private.
People have been living around village squares for thousands of years. This must have had its effect on our cultural evolution. Look at what happens when people in urban societies get to stay put for a generation or two. European cities developed structu res of plazas each with a street life suitable for a small village within the larger city. Neighborhoods developed identifications with their particular plazas. This is all considered very livable by less fortunate city dwellers from America, where peopl e rarely remain living in one place for more than ten or fifteen years.
The degree to which such a "village square" is unavailable to people is, I maintain, the degree to which people are strangers to each other, and this situation is directly related to the development of social pathologies such as criminality, alcoholism, b rutality, etc. I claim that we all have an inherent need for the function of the village square, which I call "the function of the agora".
The village square is a commons - it belongs to no one but is used by all. The agora is a commons of information - a way of interacting. It is not property.
The process of industrialization began in England with the "Enclosures Act" which deeded the village common grazing lands to whomever could build a fence around them. Needless to say, the landlords were the only ones who could raise the capital to do thi s, so the common lands went to them. This enabled them to enlarge their holdings and the resultant surplus of income over expenses provided the pool of capital upon which the process of industrialization was based. The peasants lost a source of food and were driven into dependence on what wages they could get from serving as hired labor to the landowner. All in all, a very tidy move by the landowners. Too bad about the resulting starva tion and homelessness, but, of course, there were too many peasants, anyway. The whole thing was justified because the landowners could supposedly make more efficient use of the land than the peasants.
As urbanization proceeded, a somewhat similar process of privatization of the commons of information took place. The place where people could gather and exchange information began to lose its function to the gradually centralizing mass media. There was no money in an agora which could be concentrated at some central point. But a newspaper could command a price for advertising space. So small town papers were supplanted by larger-scale publications which could underprice them. I remember vivdly the nei ghborhood shopping paper which was printed in a storefront in my native Philadelphia neighborhood. It was filled with little gossip items, each one of them of importance only to a small circle of people, but the totality of which chronicled and defined t he life of the community (a Jewish neighborhood for more than one generation). It couldn't compete for advertising dollars with larger throwaways which printed only ads and "boilerplate" generic news that was produced nationally.
All of the media had the characteristic of concentration. The only anomalies were the telephone system (but the directory was concentrated) and the postal system (which was socialized). All the rest became structured with a central point through which t he information is funneled and from which information is distributed in identical form. I call this a "broadcast" structure, and print media qualify as well as electronic media. I remember the moment in 1969 when I looked out my window down the street an d saw the living room windows glowing with the blue light of TV. I realized that they were all getting their information from Walter Cronkite in New York, but that we had no ready way to get information from each other.
In the '60's I thought that the cause of re-establishing functioning communities could be served by the newly established "underground press", and for a while I helped at the Berkeley Barb, one of the oldest such papers. But I saw the structu re of that medium determine its economics and thereby its content., and by 1970 I knew that broadcast media were never going to serve the cause of decentralization of power within society.
An encounter that year with mainframe-based network computing (through learning Basic at the SDC training facilities) alerted me to the fact that such a network had no geographical restrictions and that information items could be made accessible to variou sly defined groups of users. It was clear to me that through computer networks it would be possible to support the information needs of an overlapping set of communities of interest. But where to get a computer in those days?
Fortunately, other people had come to the same conclusions and a group named Resource One, Inc. had formed in San Francisco to secure a timesharing computer for roughly this purpose. In August of 1973 we were able to try an idea proposed by Efrem Lipkin a nd placed terminals in public places (a record store in Berkeley, followed by a branch of the SF Public Library) which people could use as a bulletin board. We called it "Community Memory".
What happened was that an agora appeared, with an unknowable number of different needs, desires, suggestions, proposals, offers, statements, poems and declarations croppping up. We, who had expected only a few categories of classified-ad items, were amaz ed at the discovery. It became clear that the crucial element was the fact that people could walk up to the terminals and use them hands-on, with no one else interposing their judgment. The computer system was not interposing itself between the individua ls who used it, either. It was serving a "secondary" information function, like the telephone directory, except that you could make your own rules as to how you were listed. When you completed your transaction on the computer, you knew who you really wan ted to talk with. The following transactions were carried out through other nonbroadcast media, mostly the phone.
The next month, September 1973, an article appeared in Radio-Electronics magazine which produced another discovery. The "TV Typewriter" was introduced - a construction project by Don Lancaster which promised that you could build a computer terminal which could display character on your TV set. The response for the mail-order plans was 50,000% above expectations - 10,000 people sent in their money where 20 had been expected.
What was going on here? Later I spoke with Lancaster and asked him why the design was not really usable as a computer terminal. He responded that "people just want to put up characters on their TV sets", and he was right. The promise of "inverting the m edia", of controlling the display of ones' own TV set, especially through a sacred-cow technology like digital computer electronics, was hard to resist. A cultural vein had been tapped.
This was the start of the hobbyist market for personal computers, from which the industry bootstrapped itself in the years 1975 - 77. Most hobbyists didn't have a good, sound, rational reason why they wanted a computer. They could make excuses about rec ipe files and checkbook programs, but nobody did that when they had a computer. They just wanted to get their hands on the technology and control it from below.
In 1978 personal computers found their first big function - communication. Alan Kay had been saying for years that a computer was first, second, and thirdly a communications tool. Ward Christensen and Randy Seuss opened a BBS in Chicago in February of th at year, and the rest is history. The number of BBS systems is unkown and probably unknowable. All this in spite of primitive software technology (of which more later). This was indeed a demand-driven application, and it is important to note that the demand was not for official, certified, top-down information, but was for contact with other people having kindred interests.
Many a well-financed "videotex" system has foundered and sunk because the operators would not consider opening the system on a person-to-person basis. People do NOT want to be subjected to centralized information. They DO want to be able to explore the s ocial space of their surroundings and to ask the question: "Who's out there?"
And then there is Internet. Like Citizen's Band radio, it is totally out of control, impossible to map accurately, and being used far beyond its original intentions. So far, so good. It has, however, developed outside a commercial structure, and if it i s placed within a commercial structure, we may expect the increasing centralization of control to which other media have fallen prey. Packet radio, developed by radio amateurs, is likewise entering the commercial arena.
Remember that the excuse for privatizing the commons was efficiency. The technology used to implement the agora function through BBS systems is actually quite primitive. The systems in use today are all derivatives of the EIES system designed by Turoff and Hiltz starting in 1971, and basic development had stopped by 1977.
These systems are message based rather than data based, having originally been conceptualized as allowing meetings with remote asynchronous participants. In such a meeting one waits one's turn or an opening while one marshals facts and figures to support a statement. When permitted by the protocol, one makes one's statement and awaits the response, ready to fend off criticism by any means necessary. This all takes place in view of all other participants and under the supervision of a moderator, who mana ges the information flow.
This is all fine for bureaucratic or academic environments, but what happens when the host stays out of the information management role and recedes into the background? The result is like having a do-it-yourself library containing not books, but scrolls, written by the people who use the library. One must take the desired scroll down from the shelf (login to a conference), read through from the beginning (there are no page numbers or indices), suffer all the frustrations of topic drift, and finally come to the end at which one may enter one's message or comment. Once written, it may not be changed. One can refer to another entry only by item number. If that item happens to be in another scroll (conference), then one may not access it without closing the scroll which is opened.
And one generally writes using an editor made by and for computer programmers. Take my word for it, they're not fun to use, or to teach. The operative attitude towards these technological shortcomings seems to be rather myopic - something along the lines of "it's good enough for me and my friends, so what's the problem?"
If it's going to survive, the agora of the future will have to be designed somewhat better from the perspective of non-technical users.
Right away, there will be objections from some quarters as to the desirability of letting in the non-techs. But they will be there whether we want them or not - the only question is whether the agora will be open or controlled for the benefit of a few. Technology matters. The personal computer took the form it did (with open architecture) because a hobbyist-based industry was free to set its own criteria without guidance from the investment community for the first two years of its life. IBM fielded the 5100, a thoroughly closed architecture machine, as soon as it could following the PC eruption in 1975. They had to withdraw that machine by 1979 and adopt open architecture in order to have impact in this marketplace. Even after they grabbed the market share with the 5150 (The IBM PC), they couldn't control the technology and close it up. The design continued to propagate like a virus through the technological body and has left them behind.
Community Memory did not dissolve after the 1973 experiment. That system was turned off in January 1975 and the people involved decided to set up their own nonprofit corporation, The Community Memory Project, in 1977 in Berkeley. Under Lipkin's technica l leadership we made a number of good calls (Unix, relational databases, X.25 as future leading technologies) and worked out a solution to the problem of system centralization through packet networking.
It took longer than anticipated, but in 1984 we put up a pilot version of the intended system at four public locations in Berkeley. In 1989, assisted by a Telecommunications Education Trust grant from the California Public Utilities Commission, we put up a ten-terminal system using a front-end/back-end architecture. We have continued to upgrade the user interface and have brought it to the level of a stable design with designed-in upgrade paths.
Most significantly, this was done without classical marketing research, but with direct involvement of a base of users over a number of years. This was "patient capital" in action. Had we gone the usual for-profit route, we would have had a product out q uickly, but it would have been indistinguishable from the others and designed with no input from users.
The Community Memory software defines a database server running on a Unix V system with 4 Mbytes of RAM, which can serve requests from PC clones running the front-end program. This program, which was developed under Unix and was ported to the PC (and is therfore portable to other systems), manages the display, keyboard, coinbox and modem. Data is exchanged with the host using an error-detecting packet protocol only when there is a data request pending. The program, which runs in 512Kbytes, buffers dat a items in local memory and allows scrolling locally without support from the server. We estimate that 50 users can access the server simultaneously without performance problems, but we have not tried a test.
The front end is based on a windowing system written for the PC which allows multiple overlapping windows. Control is accomplished by moving the cursor with the cursor keys and "clicking" using the Enter key. The display is alphanumeric using a monochro me display adapter. When the system is inactive the front end runs a "teaser" program which creates a lively animated display.
Pressing any key stops the teaser and establishes a connection to the server. The modem is dialed and login sequences are exchanged with the server (why is this not done with existing systems?). The server downloads some screens and the user is presente d with an initial screen. At any point after this a delay of a preset number of seconds without key activity on the PC will result in a warning of an impending logout followed by a logout after another delay.
The system can be navigated with the cursor and with the F1, F2 and F3 keys. F1 is always OPTIONS, which vary depending on context. F2 is always BACKUP, which will return the user to the previous operation. F3 is always HELP, which is context-sensitive. In the public implementation the keys are color-coded. The system is data based, with two types of database running. Messages (a maximum of 99 lines of text - the number is a parameter) can be stored either under an indexing system or in a network structure as comments on another message. Items are indexed a utomatically by date last edited, author, and a system-unique message ID (the tag - a six-digit alpha string). The user can select index words from a list prepared by the host of the forum (our name for conference) or may type in any amount of text, incl uding spaces, as index words. Angle brackets (<>) are used as delimiters. The index field may be edited by the user under a password protection scheme.
The ability to include text within angle brackets inside the message brings about a limited hypertext capability. One may embed references to index words, author tags, or message tags in the text, and the reader can land the cursor on them and click to e ffect a search. The next message on the reader's screen will be the message so referenced.
The reader can extend or narrow a search by subsequent search commands selected through the OPTIONS menu. At any point the reader is told how many messages were found and can scroll quickly up and down the list of found messages by use of the arrow keys. Each message occupies a window, and can be paged within the window. If a comment has been attached, a "button" is present which the reader may click to see the comment. The comment will be indented and displayed immediately below the item. Comments on comments use multiple indentation.
The hypertext capability is not formal hypertext because there is no way for a user to see who has made reference to his or her message. We have tried to create a function that the ordinary user will expect, not one which will satisfy a purist.
The commenting feature means that one can start a lateral thread and continue it indefinitely without disrupting the forum. Comment messages are not displayed until requested. People are used to handling interruptions in their conversations, and used to interrupting others' conversations. This multidimensionality is a requirement for any system which is used without an agreed-upon discipline.
Community Memory was conceptualized as a system of publication, in the sense that every message is public. We have recently added a private message (email) capability which is still in test as of this writing. Thus, when one adds a message one must resp ond to a three-entry menu asking whether the item is to be a new message, a comment, or a mail message to the author of the message which is selected.
Messages in Community Memory can be associated with any number of forums. They are not physically stored in a file associated with the forum - the connection is virtual. This association of the message to forums can be edited by the message owner or ad ded to by a forum host.
Under the coinbox rules currently in effect, it costs 25 cents to add a message and $1.00 to start a forum. These rates and structure are experimental, and with the coinbox under control of the PC, the space of possible charging strategies is large. A cr edit-balance system has been written but not implemented for dial-in use.
When we started Community Memory we purposefully did not put it on line because we wanted it not to be overwhelmed by computer types, in order to avoid a cultural barrier to the non-computer-familiar users. We took a lot of criticism from those on line for this decision, but we were rewarded by having a much more broadly representative user profile than would have otherwise been the case.
Looking ahead, we can forsee a time not too far distant in which a large number of such systems operate simultaneously, some under the auspices of institutions like the public libraries, some sponsored by businesses and by nonprofit organizations or churc hes, and all available from both private and public terminals. These systems will have the capability to be networked among servers into a centerless web to allow the transfer of messages and the automatic updating of forums which are shared among several servers. The hooks are in place for this capability now, using the protocol appropriate for Internet transport.
Through such a system one could maintain links to a number of different virtual communities and could maintain different personae as desired. One could participate in learning exchanges (teaching a skill or technique in exchange for some desired value), subscribe to various analysts and commentators from whom advice would be solicited, and engage in banter at will without disrupting things.
Through the operation of such a system, for which the technological component has already been developed, people could develop a real pluralistic grass-roots political culture without the need for stultifying meetings or the manipulation of the bored rank -and-file. Meetings in general, political or not, could be upgraded by the expectation that those who attend would have met some of the other attenders beforehand and found out something about the state of the discussion as regards salient issues.
And what it could do for trade shows and fairs is interesting indeed to contemplate. It wouldn't be utopian - nothing in systems like this would guarantee truth-telling or would protect the buyer from possible fraud, but the advantages one person has ov er another because one has channels of information and comunication and the other does not would lessen. The confidence operator, at whatever scale, who counts upon the isolation of the victims of the con would find little comfort in such a system. (Wha t would a company be like with a management information system like this?)
When I was hanging around the radical fringes of Berkeley 25 years ago we assumed that once the inconsistencies of the System were exposed and made plain, then everyone would just go about doing things the right way, whatever that was. We had only to con cern ourselves with talking about how bad things were, not how to organize things right.
Now it's recognized that as a society we're a lot closer to the crunch. The fate of the Soviet Union is staring our organizational culture in the face. The issue of "how to do things right" cannot be avoided. Pop business writers like Tom Peters are sudde nly discovering that decentralized, laterally connected enterprises function better than the centralized monsters we grew up with. The door is open both for us and the former Soviets to walk through out to the agora. But if we don't it could well close and we might follow them into a period of authoritarian reaction.
Our task, as technologists, is to build the tools that get us through this door in to the future we want. We've already done half the job by creating the personal computer such that it took on a life of its own and evaded capture. Now the task of furnish ing the agora remains. Anyone up for another adventure?
"Take the obvious...and simplify it!"