The existence of rules (formal or informal) in some sense presupposes the existence of a mechanism for enforcing those rules. The existence of such a mechanism presupposes, moreover, an ability to both detect and correct violations of the rules when they occur. The IBMPC computer conferencing facility has two basic mechanisms for the detection of problems: a formal reviewing process and the informal, but effective, vigilance of its participants, especially the forum owner. It has, moreover, four clear mechanisms for dealing with violations, including the power of expression, the reviewer and forum owner's power of erasure, the IBMPC administrator's power of denial, and management action.
When Waldbaum became the management owner of IBMPC, he recognized that the survival of the facility in IBM depended on having a reviewing process that insured conformance with the IBMPC rules and existing company standards, practices, and guidelines. This was a tricky business, as there was little precedent for understanding what would be permissible. Each append and forum had to be considered carefully in the facility's early years, and the formal rules of the facility evolved iteratively in response to the problems that were encountered. Fortunately volume was fairly small between 1981 and 1983, usually amounting to only a few contributions a day. It was possible for any participant to read every contribution to the facility in only a few minutes a day, and it was possible for the administrators of IBMPC to think carefully about each one.
The process then, as now, was simple. Contributors made appends to IBMPC. The appends were distributed immediately and reviewed later. If there was a problem, the append was revised or, if necessary, deleted altogether. Waldbaum gave Chess, a programmer, responsibility for reviewing, developing, and running the facility. Individual forum owners had been given (in the TOOLS commands) the ability to administer their own forums.
The decision to "post-review" appends was, by contrast, a difficult one. It would have been easier, from the perspective of managing a new resource, to require the "pre-review" of all appends before distribution, as it would have insured that every contribution distributed on IBMPC would conform to the IBMPC rules. There were, however, at least two problems associated with pre-review. First, it could be expected to put considerable pressure on Waldbaum and his staff, as submitters would want their submissions posted quickly, regardless of when they were sent. Second, it was expected that the inevitable delay in posting submissions would have an inhibiting effect on the interpersonal character of computer conferencing.
It was (and is) felt that pre-review of appends inhibits conferencing by slowing things down. Even if one could monitor incoming appends continuously and process them immediately, pre-review would delay appends by at least several minutes, effectively reducing the conversational character of conferencing. It could reasonably be expected, moreover, given available staffing and competing responsibilities, that many appends to IBMPC would be delayed by hours or even days (weekends, holidays, etc.) in a pre-review process. The effect, it was felt, would be to eliminate the possibility of anything close to real time interaction in computer conferencing.
The decision to review appends after they have been posted to the conference represented, then, an attempt to shape the rules of IBMPC to the medium. It might have been easier to shape IBMPC to conform to a more restrictive and less risky set of rules, but IBMPC probably would not have been as successful if a pre-review system had been used.
There's no record or recollection of when the first problem requiring administrative action occurred, but the process for detecting the problem was fairly simple. Genuinely interested in personal computers and what could be done with them, Chess, as a participant on IBMPC, read (or tried, for software) everything contributed for at least the first three years of IBMPC's existence. When he saw what might be a problem (or had a potential problem pointed out by a participant) he deliberated with Waldbaum on what action should be taken. In many cases, IBMPC's formal rules were revised appropriately.
Chess wasn't alone in participating on IBMPC. Waldbaum also participated from the beginning. Later, others on his staff, including at least one other PC consultant and several managers, also read IBMPC regularly. Before 1984, Waldbaum took over if Chess was unable to review the facility. After 1984, responsibility for IBMPC was sometimes delegated to others for short periods of time.
With the more rapid growth in contributions to IBMPC that started in late 1984, it became increasingly apparent that the task of reviewing IBMPC would have to be shared on an ongoing basis. Hence in early 1985, the job of reviewing IBMPC was split up between several people, including Chess, Waldbaum, John Alvord, Henry Morse, and a PC consultant, with each responsible for following a subset of the forums on the facility. This solution also proved inadequate to the accelerating growth in contributions. Waldbaum, the management owner of IBMPC frankly acknowledged the need for a better solution in a May, 1985 append to SENSITIV FORUM: "IBMPC needs a full-time reviewer."
Viewed purely from the perspective of workload, the arrival of Gloria Whittico, IBMPC's first full-time reviewer, five months later solved some problems. The new reviewer, a former IBM attorney, was able to take over all the formal responsibility for finding and resolving problems on IBMPC in fairly short order, leaving IBMPC's other administrators with more time to do other things. Whittico was also asked to play what may have been an even more important role, however, in working to institutionalize computer conferencing in IBM, first by selling the idea of conferencing to IBM executives, and later by working to help create IBM's corporate computer conferencing guidelines.
The arrival of the new reviewer, the first for whom reviewing IBMPC was a full time job, also raised a set of challenges. The new reviewer was not, for instance, an IBMPC participant and could not be expected to be familiar with the either IBMPC's informal rules or the way IBMPC's formal rules were interpreted. The new reviewer's background in IBM legal work resulted, moreover, in new perspectives on what the formal rules ought to be and how they ought to be interpreted. Hence such issues as what constitutes a PC related discussion, what constitutes a business related discussion, what constitutes professional behavior on a computer conferencing facility, and what the role of the IBMPC reviewer should be were brought into sharper focus, forcing IBMPC's administrators to clarify many of the boundary conditions of IBMPC's rules. A number of these issues are discussed below.
The most difficult issue faced by the IBMPC reviewer is the the question of what can be discussed on the facility. This issue is a recurring one that has sparked debate repeatedly over the course of IBMPC's history. Although the IBMPC rules impose a general constraint that appends be both IBM business and PC related, the implementation of this general constraint has been problematic for two reasons.
First, it is difficult to obtain general agreement on what constitutes a definition of PC or IBM business relatedness. The question of PC and IBM business relatedness is not, of course, a matter that is decided by general consensus. The only opinion that ultimately matters is that of the manager owner of IBMPC and the management he is responsible to. Lack of general agreement, even in the face of such decisions, can create continuing problems, however, as individual employees and managers, often working from different perspectives and assumptions, can and do periodically question such decisions.
Second, and perhaps more importantly, it is difficult for any definition to account for all the potential boundary conditions. Hence new borderline cases force the administrators of IBMPC to continually refine the IBMPC rules and their interpretations of the rules.
In defining business related, for instance, some would assert that business related means directly related to a product that IBM currently produces. Others would assert that business related means related to any product that IBM might be reasonably expected to produce given its current lines of business and general market trends. An even more liberal view might substitute the word conceivably for the word reasonably. None of these definitions is entirely adequate to business relatedness, however, as many of the kinds of knowledge that are required to invent, plan, design, program, manufacture, test, service, sell, and otherwise support a product are not directly related to any particular product, even when they are indirectly related to all of them.
A similar varied array of definitions can be constructed for PC relatedness. A restrictive definition might restrict discussion to IBM's PC products. A less restrictive definition might allow discussion of any product that can be used with IBM's PC products. An even less restrictive definition might allow discussion of any PC product, whether IBM's or not. A somewhat liberal definition might allow discussion of any use to which a PC might reasonably be put. A more liberal definition might allow discussion of any use to which a PC might conceivably be put. An extremely liberal definition might allow discussion of any topic that personal computer users are generally interested in. Again, none of these definitions will be entirely adequate, as each misses large areas of critical interest to people who are creating, marketing, and using PC products.
Although there have been advocates for each of these positions (and others) over the history of IBMPC, the IBMPC administrators have been fairly persistent in using a definition somewhere in the range between "reasonably" and "conceivably" as its broad criterion of both business and PC relatedness. The real problem has never been the definitions, which have changed very little since moving somewhat closer to "reasonably" after the full-time reviewer was brought on. The problems, for the most part, have entailed the question of boundary conditions.
Consider, for instance, a meta-forum like SENSITIV FORUM. It clearly is not PC related. The only argument that can be made for its business relatedness is that it helps keep the peace in other forums. Consider, for contrast, BALLISTC FORUM, a recently closed conference that started as a discussion of ballistic algorithms and programming on PC's. One can easily imagine IBM selling PC products that use ballistic trajectory calculations. Indeed, IBM probably has sold game and other software that used such algorithms. For a third example, consider SPACE FORUM, a fairly popular conference that, in theory, discusses the use and potential applications of PC's in the exploration of space. If used as intended, this forum is both PC and business related. IBM sells computers and software to the U.S. space program. At least some of those computers are PC's. At least some of that software runs on PC's.
Two of these forums, SENSITIV and SPACE, are ongoing and long standing conferences on IBMPC. One, BALLISTC, was fairly short lived and is now closed. Two of these forums, SPACE and BALLISTC, have clear PC relatedness and product relevance. All are probably borderline, however. None would probably survive an extremely narrow definition of PC and business relatedness. If one were to draw a (clearly subjective) continuum, BALLISTC FORUM would occupy the middle ground among these forums. Why, then, is it closed while the others continue?
If one scanned BALLISTC FORUM looking for an answer, one might well conclude that, despite good intentions, it was never able to stay on topic. This conclusion is a fairly accurate one. BALLISTC FORUM was filled with appends discussing all sorts of ballistic esoterica, including discussions of barrel rifling, the National Rifle Association, and preferred weaponage. Almost none of these appends was in any way useful to anyone interested in making use of a PC for the purposes of doing ballistic calculations. It was ultimately closed by the IBMPC administrators because of this propensity, its resistance to reviewer suggestions that it get on target and persistent complaints from the IBMPC community about the forum's lack of business relatedness.
Many of the same things can be said, however, of SPACE FORUM. SPACE FORUM attracts a variety of appends about launch schedules, characteristics of various planets, alternative systems for putting objects in space, etc. The forum often grows quickly. When a space shuttle is preparing for launch or is in orbit there may be dozens of appends in a day. Clearly the ability of a forum to stay on topic does not fully explain why one forum survives and the other is banned.
This is precisely the problem of boundary conditions. The rules do not, and cannot, exactly describe the border between what is and is not an acceptable topic for discussion. Decisions at the boundaries hinge on other more arbitrary factors, including what people are interested in, how much work different decision options entail, and management support. SPACE FORUM survives BALLISTC FORUM for exactly these kinds of reasons.
Interest in BALLISTC FORUM was fairly narrow. Participation levels were low. Several months could elapse with no appends. Readership, as measured by subscriptions, was fairly small. When the forum was challenged, its readers offered few justifications for the forums continued existence. Interest in SPACE FORUM, by contrast, is extensive. The forum has many readers and a large number of contributors. This interest reflects, in part, IBM's participation in the U. S. space program. More important, however, is its reflection of a basic motivation behind many IBM employees participation in the computer industry. This motivation, the belief that the exploration of space is an ultimate goal of developing computer technology, will be reflected strongly as we discuss the short lived SHUTTLE FORUM.
The work entailed in maintaining BALLISTC FORUM, including erasures of appends and responses to complaints, was fairly large relative to its apparent value. The IBMPC administrators were required to take some action related to BALLISTC every few months. SPACE FORUM, by contrast, has generated very little work for the IBM administrators. Most problems that occur on the forum are cleared up by the forum owner or forum participants before the IBMPC administrator has a chance to become involved. Where administrator action has been taken, moreover, it has generated little controversy.
The IBMPC administrators remain confident, moreover, that the participants in SPACE FORUM would be able to build a strong case for the forum's continued existence if asked to justify the forum's continued existence. SPACE FORUM is the primary vehicle for information exchange between IBM employees who work in the space program and those who work on other projects for IBM. While the benefits of this exchange have not, as yet, been enumerated, it seems fairly certain that they include transfers of technology to and from the space program that would not have happened otherwise.
The IBMPC community has always been interested enough in SPACE FORUM to put in the extra work that the forum entailed. BALLISTC, by contrast, was more work to the IBMPC administrators than it seemed to be worth. Hence after failing to come up with a formula that would make the forum worthwhile, the IBMPC administrators gave up on it. SENSITIV FORUM and other meta-forums, although related only to the business of maintaining order on the conferencing facility, exist because the IBMPC reviewers believe they substantially reduce the incidence of problems on IBMPC.
Decisions concerning the content of IBMPC are usually more clear-cut than has been illustrated here. Most forums on IBMPC are clearly both PC and business related. Most of the appends that are removed from IBMPC by the IBMPC reviewers are clearly neither. It is the boundary conditions that give the IBMPC reviewer the most problems, for they are neither easy to resolve nor, once resolved, to explain. If asked why BALLISTC FORUM is banned, an IBMPC administrator is most likely to say that the forum wasn't sufficiently business related. The actual reasons are somewhat more complex, and much more difficult to state.
Such boundary conditions are inevitable, regardless of where one draws the line between business related on non-business related. Thus so long as one regards continuity as a goal in such decisions, the existence of such difficult boundary conditions will present a challenge to new IBMPC reviewers, especially when they are not participants on IBMPC to begin with. The training of a new IBMPC reviewer is, then, more than just an education in the rules of IBMPC. It is an education in IBMPC's boundary conditions that can only be taught by example. The education of a new reviewer begins by turning the reviewer into an IBMPC participant. The reviewer starts by simply reading IBMPC, with particular attention given to borderline forums. Reading is an appropriate start, as most of the IBMPC reviewer's job is reading. The more problematic forums are an appropriate place to start, as IBMPC's boundary conditions will be the biggest problem the IBMPC reviewer has to deal with.
Perhaps the most obvious problem that had to be addressed after IBMPC's first full-time reviewer arrived was the question of volume. The volume of appends to IBMPC was daunting in late 1985, rapidly approaching the volume of a fair-sized novel each week, and the new IBMPC reviewer rapidly accumulated printouts of new appends which sometimes piled waist high from the floor. Growth has continued, moreover. In late 1989, reading all of IBMPC adds up to about 460 single spaced pages a day. Hence nobody, including the IBMPC administrators, reads everything appended to the facility anymore. The reviewing process depends instead on a combination of append-monitoring software and a watchful IBMPC conferencing community.
This append-monitoring software can be thought of as a kind of automated IBMPC reviewer. It scans every new contribution to IBMPC and selects a subset of those contributions for review. Several criteria are used in selecting appends:
Some of the key words on this list can be regarded as permanent members. This list includes several English language obscenities and other words including "copyright", "confidential", "reviewer", and "encrypt". Appends containing the word "confidential" are a matter of concern because the may discuss confidential information (or appends that may discuss confidential information). The word "copyright" sends out a similar signal insofar as the copyrighted material of others should not be posted to IBMPC. "Encrypt" is included because of U.S. government restrictions on the dissemination of certain encryption algorithms. "Reviewer" is included because any discussion that references an IBMPC reviewer probably entails meta-discussion and/or a problem.
The list can vary somewhat, depending on recent events. The words "stock" and "crash" joined the list of key words after the October, 1987 stock market crash, and disappeared a few weeks later. The names of the presidential candidates and the words "campaign" and "election" joined the list of key words during the 1988 presidential election, and disappeared a few days after the election was over. The names of IBMPC administrators and reviewers who change jobs are usually added to the list for a short while after they give up their IBMPC-related responsibilities.
One such forum (recently closed) was NEWSCHAT forum (and its predecessor, NEWSTALK FORUM). NEWSCHAT FORUM was (as NEWSTALK was) a forum where news items that have been appended to NEWSCLIP FORUM could be discussed. The content of NEWSCHAT was, by definition, as broad as the range of news reports that relate to the PC market and IBM's participation in it. This content is also as explosive as the range of peoples opinions about that news and what it means to IBM. The rules of NEWSCHAT FORUM (a variant of the NEWSCLIP FORUM rules) sought to reduce this explosiveness by declaring all of the following as off limits:
The last of these restrictions is a catalog of topics that have repeatedly attracted a great deal of comment in a short period of time without demonstrating any PC relatedness. It was the tendency to this kind of discussion that repeatedly earned NEWSCHAT a prominent place among the small list of forums for which every append was reviewed. These same propensities eventually lead to the forums closing.
A less pressing, but ultimately unavoidable, issue asks what it means to review IBMPC. The reviewer of IBMPC is filling a role and the definition of that role will determine how they perceive their role, what they do in fulfilling that role, and how other people perceive their performance of that role. The reality of this role is fixed by its job description. The IBMPC reviewing process is a business process which is very similar to the role IBM management takes on when it reviews a document for external distribution or reviews a bulletin board to insure that its contents are business related. The IBMPC reviewer is a management designee who has the responsibility of insuring that contributions to the facility meet its rules.
The nature of this specific role is well understood by IBM managers. Non-manager participants on IBMPC often understand this role as analogous to that of an editor, and sometimes as analogous to that of a censor or police officer. Many non-management IBMPC participants view the IBMPC reviewer's role as analogous to that of an editor, reviewing prospective forum content, accepting material that fits within IBMPC's editorial policy, and editing borderline material to meet the facilities needs. Frustrated prospective appenders sometimes react to reviewer actions by charging censorship. Some, moreover, regard the reviewer as as a police officer who searches for and punishes behavior that, according to the IBMPC rules, is illegal.
The job of editor is one of reviewing copy, assessing how well each contribution fits the editorial objectives of the publication as expressed in its editorial guidelines (in this case the IBMPC rules). When an editor rejects copy, the rejection is not a statement that it cannot be published anywhere. It is simply a statement that, as written, it doesn't fit a particular publication's editorial needs and objectives. When an editor modifies copy, it isn't a statement that something isn't written well. It is a statement that it better fits the needs and objectives of the publication with these changes.
If viewed from this perspective, IBMPC is a publication to which participants contribute copy. The role of the IBMPC reviewer, if acting as an editor, would be to control the content of this publication by accepting or rejecting copy and perhaps editing some portion of the accepted copy to meet IBMPC's editorial needs. As an editor, the IBMPC reviewer would have some latitude in decision making. An append that didn't quite fit within the editorial guidelines but that was still of considerable interest to the readership might be accepted. Similarly, a series of appends that is perfectly within the rules but which addresses a topic in the same way for the fifteenth time might be cut off with a back pointer.
If there is a weakness in this view, it is the relatively limited amount of "editing" that the IBMPC reviewer actually does. The reviewer reads a lot, anywhere from 10% to 50% or submissions, but only "edits" a very small fraction of appends. Indeed, true editing is the exception. It is not the practice of the IBMPC reviewers to change people's words. Borderline rejections may be accompanied by suggestions about how the contributor might edit and resubmit the append. Borderline acceptances may still entail a note to the appender suggesting changes that would improve the append. Existing appends are simply accepted or rejected, however.
Contributors to IBMPC appear, moreover, to have an excellent understanding of the scope of the facility and its rules. Hence very few appends, only about 1 in 650 according to a recent analysis of the contents of IBMPC, are rejected.
The view of IBMPC reviewer as censor is not as simple, in part because the role of censor is somewhat ambiguous. The television censor, for instance, acts as a "guardian of public morals". The government censor, by contrast, acts to insure that formal restrictions on what can and cannot be expressed are enforced. The effect of both censors is the essentially the same. Freedom of expression is restricted. The scope of the effects are clearly different, however.
At its extremes, government censorship acts to restrict all expression through all media on defined topics and/or from defined points of view. What cannot be published in one magazine cannot be published in any magazine or, for that matter, even written in a letter. What cannot be said cannot be said anywhere. The television censor, by contrast, acts to restrict certain forms of expression from a single instance of a single media. What cannot be expressed on one channel or at a particular time of day may well be allowed on another channel or at another time of day. The actions of government censors may, then, be considerably more constraining than those of the television censor.
Indeed, the actions of any given television censor are similar, in many respects, to those of an editor. If there is a difference between editor and television censor, it is in the nature of the restrictions imposed and the breadth with which those restrictions are applied across instances and media. The television censor is viewed as a censor rather than an editor in part because there is a censor applying almost identical standards for every instance of the medium. The effect of this application of a single standard across instances is a broad based censorship. The broad application of a singular set of restrictions simply makes it impossible to express certain ideas on television.
The restrictions applied by an editor are, by contrast, much less constraining. Any given publication can be expected to constrain the range of ideas explored in its pages. Most publications may, moreover, enforce similar standards for spelling and grammar. There is probably no idea, however, that cannot be expressed in some broadly distributed newspaper or magazine.
The restrictions applied by the IBMPC reviewer are similarly unconstraining. The IBMPC reviewer makes decisions for a single instance (IBMPC) of a single medium (computer conferencing). The topics that are constrained on IBMPC are not, moreover, constrained by a global ban on the topic. While, for instance, IBMPC discourages legal discussions, many IBMPC participants may have job responsibilities that entail some measure of legal discussion. They simply can't talk about such issues on IBMPC, where geographically dispersed participation makes it impossible for specific legal opinions to be of any value to most participants.
The primary restrictions on what can be expressed on IBMPC state that contributions should be PC and business related. Neither restriction can be regarded as being broadly applied across other computer conferencing facilities. There are many computer conferencing facilities outside IBM that will allow anyone, including IBM employees, to express ideas that are not business related in IBM's view. Many of these facilities have few restraints on the way those ideas are expressed. There are, moreover, many computer conferencing facilities inside IBM that allow IBM employees to explore business related ideas that are not PC related.
Reviewing on IBMPC differs from both censorship and editing in one other key area. Editing is something one does before publication. Censorship is something one does before a message is broadcast, published, or otherwise transmitted. Hence both acts prevent deleted or modified content from reaching an audience. Both acts, when done well, can be invisible to an audience which may never realize content was deleted or changed.
Reviewing on IBMPC, by contrast, is generally done after an append is broadly distributed. Many IBMPC participants may view an append before any reviewing action is taken. When taken, moreover, reviewing actions are visible as modified appends. A line at the top of the append indicates that the append has been modified and who made the change. Hence an action taken by an IBMPC reviewers tends to be fairly obvious to anyone who sees it.
The inherent visibility of this reviewing process sets up a natural accountability for reviewing actions that may not be present in other media. Neither the editor nor the censor are generally highly accountable to their medium's audience for their actions. A newspaper readership may never know that an editor suppressed some significant element of a news story at police request. It might be years before a novel-reading public learns that an important novel is strikingly different than a preserved original manuscript. A television audience may never realize that a scene was softened at the request of censors. Every reviewing action on IBMPC is visible, however.
Another analogy equates the IBMPC reviewer to a police officer charged with upholding the law and keeping the peace. The police officer really has two roles. One is that of keeping the peace; breaking up disputes that are getting out of hand; preventing small problems from turning into big ones. The other is investigating violations of law and taking action to punish violators. When viewed within this analogy, the job of the IBMPC reviewer is to patrol the forums of the IBMPC computer conferencing facility looking for potential trouble. Where one sees a problem developing one may act to keep things from getting worse by putting in a conciliatory append or asking participants to back off. Where one sees a violation of the IBMPC rules, one acts to resolve the violation by removing the append or asking the appender to change it.
If there is a strength to this analogy, it is its essential post-review perspective. The police officer rarely gets to anticipate events and stop a crime before it happens. More often the job is, like that of the IBMPC reviewer, one of reaction to events that have already occurred. The weaknesses in the analogy are considerable, however. Where the police officer would punish the individual who violates the rules, the IBMPC reviewer often simply undoes the violation. Indeed, so long as the problem is not so severe as to demand immediate action, the reviewer will most often seek to persuade the contributor to correct the problem.
The value of these analogies is limited. Thinking of the IBMPC reviewer as an editor may help IBM employees who are unfamiliar with IBM management processes to understand the role, but does not provide an accurate view of what the IBMPC reviewer does. Invoking such terms as "censor" or "thought police" may have some rhetorical value (who likes being called a censor), but the rhetoric does not reflect the reality.
The actual role of the IBMPC reviewer is unique to the computer conferencing medium. It represents, from the standpoint of both media structure and IBM management practice, a new and distinctive role which may someday itself be used as an analogy for understanding similar roles.
Viewed from the standpoint of media structure, the role of the IBMPC reviewer has evolved to fit the needs of a distinctive medium. The combination of interpersonal and mass media characteristics that characterize computer conferencing place demands on the role of the IBMPC reviewer are different than those associated with mass media. Interpersonal content is distributed to a mass audience, and the interpersonal elements of the medium are compromised by the kinds of controls that are normally used to constrain content in mass media.
Viewed from the standpoint of IBM management practice, IBMPC and computer conferencing offer the company possibilities that simply have not been available in other media. IBM has traditionally dealt with interpersonal communication practice through its business conduct guidelines, which proscribe the limits of personal behavior when one is representing the company. IBM has traditionally dealt with mass communication practice through document clearance procedures in which documents (including this one) are carefully reviewed to insure that the interests of the company are protected. Computer conferencing, which has characteristics of both, requires more sophisticated controls that combine elements of both its interpersonal and mass media controls.
The role of the IBMPC reviewer has evolved with the IBMPC rules to fit that need. The IBMPC rules provide the conference specific equivalent of the IBM Business Conduct Guidelines. Participants are expected to be familiar with the IBMPC RULES and to contribute in a manner that is consistent with those guidelines. The IBMPC reviewer's formal responsibility, relative to these expectations, is to ensure that the rules really are observed. This sometimes means deleting an append, a thread of discussion, or an entire forum. Occasionally it entails more extreme measures which will be discussed later in this chapter. More often, however, it means communicating with individual appenders and the IBMPC user community. The IBMPC RULES are a part of this communication process. Meta-forums like REVIEWNG FORUM and SENSITIV FORUM also provide a critical means of communicating these expectations to the IBMPC user community.
The IBMPC reviewer's most powerful tool is probably his or her ability to show why a recommended course of action is the right one to take. The act of deleting an append may or not be effective over the long term. The act of persuading the IBMPC community that a given type of append only leads to problems may mean that similar appends appear less often, and are then often handled without the IBMPC reviewers intervention. An appender can often be convinced to change a problem submission through the simple expedient of pointing out the problem. A strong disagreement between participants can frequently be dealt with by pointing out the common ground in their otherwise contrasting positions. A problem discussion thread can often be stopped or redirected with a few well chosen words. Hence the IBMPC reviewing process is basically one of communication and meta-communication.
It is this process of communication and meta-communication that most strongly differentiates the role of reviewer from other analogous roles. The censor prevents. The editor rejects. The police officer enforces. The reviewer, when at his or her best, most often persuades.
Such persuasion is hardly restricted to the IBMPC reviewer, however. Most regular IBMPC participants have a fairly reasonable understanding of both the formal and informal rules governing IBMPC. Indeed, many have, at one time or another, helped to refine these rules. Quite a few have come to better understand those rules in the process of breaking them. All evidence indicates that they are dedicated employees of IBM who understand that the rules are there to protect both the company and IBMPC's participants.
It should not be surprising, then, that IBMPC participants have never been shy about pointing out what they perceive to be problems, usually with a note sent to the appender and copies sent to the forum owner and/or IBMPC administrators. Where the problem is clear cut and obvious, IBMPC's administrators often remain uninvolved. The referenced append is frequently modified by the appender before the IBMPC reviewers even know there is a problem. Some IBMPC participants are bothered by the actions of these "self-appointed reviewers", but this derisive label misses the point.
IBMPC's informal reviewing process is as inevitable to the medium as smiles and frowns are in face to face interaction. The note that asks an appender "Do you think you made a good choice of words in your append to MARKETNG FORUM?" is, in effect, a frown. So is the note that states that "I found the language of your append to MARKETNG FORUM offensive. I think the append should be deleted." Both notes tell the appender that the reader found the append wanting. Both notes improve on a simple frown, moreover, by suggesting what can be done to improve the referenced append.
These notes are acts of meta-communication. They attempt to propose mutually acceptable standards for interaction on the medium. The effect of the notes is to exert peer pressure for conformity. One note may be more subtle than the other. One note will probably provoke a better response than the other. There is no guarantee that this kind of informal action will be successful. Success is contingent on the critic's ability to convince the contributor that there really is a problem that ought to be addressed. Such notes work best in those cases where an append was written quickly and the author didn't realize how some part of the append might be read. Modifications of appends usually appear quickly in such cases.
Things get more complicated when the appender doesn't agree with the assessment of the critic. It is frequent, in such cases, for the contributor and critic(s) to exchange a series of notes, often without any resolution beyond agreeing to disagree. It is also fairly common for one party or the other to involve other opinions, including those of the IBMPC administrators and, in some cases, participants in IBMPC's meta-forums. The selection between these resources is, to a large extent, contingent on the problem.
Perceived violations of IBMPC's formal rules will most commonly result in an appeal to the IBMPC designated reviewers. IBMPC participants are, for the most part, fairly careful with their criticisms of appends. Hence when the IBMPC reviewers receive a complaint it is examined carefully under the presumption that there probably is a problem to be resolved. The decision of the IBMPC reviewer, in this context, is generally final, in part because the reviewer has the power to make a decision stick (more on this later).
Still, there is no guarantee that the reviewer will agree with the critic. Where a critic charges that an append is not business related, for instance, the IBMPC reviewers and management owner may not agree. Where a critic charges that a particular word is obscene, the IBMPC reviewers will sometimes disagree because the word does not appear to be used in an offensive manner.
Resolution of violations of IBMPC's informal rules often hinges on the independent actions of many members of the IBMPC community. In the simplest case, a large number of people see a problem in an append and write to the author about it. It is not unusual, in such cases, for a change to be made even when the appender disagrees. Peer pressure can be powerful when it starts to fill up one's electronic mailbox.
A second major approach is found in meta-forums. The meta-forum acts, in such cases, as an informal town meeting of IBMPC participants where the propriety of a given action can be discussed. The power of such town meetings should not be underestimated. There are, of course, issues that have never been (and probably will never be) fully resolved in these meta-discussions. There are, moreover, individuals who will persistently violate the informal rules of IBMPC as decided in such forums, even when they have participated in the discussions.
The power of such discussions seems to be considerable, however. Although this February, 1986 append to SENSITIV FORUM may overstate the case somewhat, it does express the kind of effect meta-forums like SENSITIV FORUM have had on the IBMPC computer conferencing community:
I think the absence of appends to this forum in the past month and a half speaks loudly for increased sensitivity of (us) conferencers to each other. I've been following several more discussions ... than I used to, and even so I've noticed a surprising lack of:despite the continued presence of
- snide comments
- complaints from offended parties
- apologies to offended parties
- deletions of appends
Are we really learning the language of green (or varicolored) phosphor dots as well as this suggests? I hope so!
- stupid questions (hooray for stupid questions!)
- erroneous answers (boo hiss)
The append is just one person's opinion, but few IBMPC participants would disagree with its basic sentiment. Meta-forums have been very successful in allowing IBMPC participants to work out their rules of engagement. A great deal of the character of communication on IBMPC results directly from the preferences of its participants as expressed in meta-forums and private notes.
These informal actions are all expressions of a general power which might be called the "power of expression". Persuasion, in the form of private notes between participants and appends to meta-forums, is the tool of choice when resolving perceived violations of IBMPC's formal and informal rules. It does not, of course, always work, but it is likely, when it does work, that everyone will leave the field of battle satisfied, if not entirely happy, with the result.
This is at least as true for the administrators of IBMPC as it is for other participants. The IBMPC reviewer's tool of first choice for resolving problems which are not serious remains a note to the contributor. Sometimes an append is such that it requires immediate action. More often, however, solution of the problem is not particularly time critical. Under such circumstances the IBMPC reviewer is most likely to write a short note to the appender that outlines the problem and an acceptable solution.
The same is true when the line of discussion in a forum either overheats or takes off on a tangent that is neither PC nor business related. An append by an IBMPC administrator that attempts to redirect discussion is usually effective in bringing discussion back into line. Where it isn't, a formal warning append almost always solves the problem.
If a note or append is not enough, either because the content of an append and/or forum must be dealt with immediately or the appender(s) aren't listening, stronger tools, in the guise of the "power of erasure", are available. In the case of a single problematic append, two specific actions are available. First, the reviewer may edit the append, either removing or changing a problematic passage in an otherwise acceptable append. This is not a preferred option, as the original appender may not be happy with the change and its effect on the rest of the append. It is, however, sometimes exercised when, aside from the problem passage, the append seems important to the flow of the forum. Second, the append can be deleted from the forum entirely.
These actions represent fairly extreme measures. There are a fair range of cases which justify such action, however, including:
When an append is edited or erased, a note that details the problem is sent to the contributor. The note includes the original append, which the contributor is free to resubmit to the forum so long as the problem is resolved. Many contributors make such changes and replace the reviewer's deletion with a personally revised version. Hence while the power of erasure is substantial, it is not absolute.
The rules change, however, when a whole series of appends in a given forum need to be revised or deleted. Such action, which the IBMPC administrators sometimes jokingly call an appendectomy, is usually only required when a discussion gets out of hand before anybody notices it. The procedure is an unusual one (the IBMPC administrators rarely needs to take such action more than once over a period of several months). When such action is necessary, 10 or more appends will often require removal, and the sheer volume of the effort changes what the reviewer can reasonably accomplish.
If a note is sent to the appenders whose appends are deleted, it will generally be a single note sent to all. The deleted appends generally will not be attached. It is often the case, however, that the note will be replaced by a single brief append, with any discussion of the action referred to REVIEWNG FORUM. Depending on how the deletion is handled, moreover, it may not be possible for contributors to replace their old append with an acceptable one. If the appends are deleted individually via append-modify, replacement will be possible (although it still may be discouraged). If, however, the forum is pruned or replaced entirely with an edited copy, it will not be possible to replace the deleted entry with a modified version of the append.
Forums that attract frequent attention from the IBMPC administrators may be subject to other actions, including renaming or closure. Consider SCIFI FORUM, for instance. The formal purpose of SCIFI FORUM was the discussion of the use of computers in science fiction and the technical and marketing possibilities suggested in such use. When on track, the forum frequently discussed some of the wild applications of computers which could be envisioned. It was also home to interesting discussions of how, in personal computers and other technologies, science fiction was becoming contemporary reality. If SCIFI FORUM participants had been able to restrict themselves to these and similar discussions involving the intersection of science fiction and computing, the forum would probably still be operating.
A single propensity of the forum seemed to make this impossible, however. Participants insisted on periodically ignoring the issue of computing and simply discussing their favorite science fiction. The pattern associated with these digressions was a fairly natural and rather innocuous one. One contributor would make a perfectly reasonable append about the use of computers in a work of science fiction they had just finished. Another appender would then reply with a comment about how good or bad the book was that made no reference to computing. Once the second append was in, the direction of discussion was, if not set, clearly pointed away from the use of computing in science fiction. Subsequent appends might:
Viewed from the perspective of literary review and discussion, these tangents were often extremely interesting and well thought-out. Viewed from the perspective of business and PC relatedness, the book review discussions were completely out of bounds. They frequently choked out the valid discussion that did occur on the forum. Reviewer warnings and appendectomys were a commonplace event through most of SCIFI FORUM's history. With IBMPC's continuing growth, moreover, actions against SCIFI became more frequent. Eventually, despite the unquestioned value of the forum when it remained on target, the reviewing workload and misuse of resources (including disk space, network resources, and the time of IBM employees) became unacceptable. The idea of SCIFI seemed a good one. Science fiction has a great deal to say about the future of computing. But the forum wasn't working.
Feeling that the forum's name might be a big piece of the problem, the IBMPC reviewers tried giving it a name that more closely reflected the purpose of the forum. Hence SCIFI FORUM became SFIDEAS FORUM. The theory, in renaming a forum, is that participants judge the appropriate content of a forum, at least in part, on the basis of its name. In the case of SFIDEAS, IBMPC's administrators tried to de-emphasize science fiction relative to the computing related ideas found in science fiction. Other names, including SCIFI-PC, SCFUTURE, or SFCOMPUT, might, in retrospect, have been better choices, but SFIDEAS caught the general notion that this was a forum for the discussion of the ideas in science fiction rather science fiction itself. Forum renaming has been tried several times on IBMPC, often with success.
It appeared, for a short time, to have worked for SFIDEAS as well. With time, however, the old pattern crept in. With new forums, including the similarly named FUTUREPC and PCFUTURE FORUM filling a similar niche, but without the problems, the workload associated with SFIDEAS became more difficult to support. It was eventually closed.
The closing of a forum is a first example of the "power of denial". Denial is, in all its variants, a much more extreme solution than erasure. Erasure makes a statement that a contributor or group of contributors made a mistake, but preserves the ability of the contributor to learn and avoid the mistake in the future. Denial, in the case of a forum closing, makes a stronger statement concerning the possibility of conducting consistently business related discussion on that forum. It denies the possibility of making the same mistake again.
At least six variants of denial are available to the administrators of IBMPC:
Although two cases of forum closings have been described in this chapter, the event is an extremely unusual one. Well over 5000 different forums have been created on IBMPC over the history of the conferencing facility. The number of active forums that have been closed by the IBMPC administrators is probably less than 150. Most of these were closed, moreover, either because they were redundant with an already-existing forum or they were not adequately PC or business related.
The TOOLS "LOCK" command allows a forum owner or IBMPC administrator to close a forum to subscriptions without deleting its content from IBMPC. When, for instance, one is about to perform an appendectomy on an out of control discussion, locking a forum helps to insure that newly inflammatory appends don't have a chance to undermine the effort. Participants can still read the forum. They just can't contribute to it. The act of locking a forum is generally a very temporary one, lasting only a few minutes or hours.
The act of putting a forum on preview is a less extreme action. Forum preview has historically been used to insure that new contributions adhered to the scope of the forum and met the IBMPC rules. It has frequently given an overheated forum an opportunity to quiet down. Preview is more work for the IBMPC reviewers, who must review appends as they are made if interaction is to proceed smoothly. It is, moreover, frustrating to contributors whose appends are sometimes substantially delayed by the process. It does, however, allow problem appends to be handled without them first being broadcast to large numbers of individuals. Events requiring a preview action don't happen often. Indeed, several months may pass between forum preview actions. Its use is, moreover, almost always fleeting. Forum preview actions rarely last more than 24 hours.
The act of putting an IBMPC participant on preview represents a somewhat more serious power of denial. It is reserved for individuals who repeatedly or seriously violate the IBMPC rules. All appends made to IBMPC by an individual that has been placed on preview are reviewed before being posted to a forum. There are very few instances of a user being placed on preview. Where such action is taken, however, it is usual for the preview to be maintained for weeks or months.
The least extreme, but frequently most controversial, form of denial is the closing of a topic to further discussion. This form of denial is not uncommon on IBMPC. It is, in fact, implicit to every erasure. This does not mean that there are as many closed topics as there are erased appends. Most append erasures are related to a small set of restrictions, including business and PC relatedness, that are documented in the IBMPC rules. Hence erasures are much more likely to reflect an existing sphere of denial than they are to reveal a new sphere of denial.
Sometimes, however, it is necessary to close an otherwise reasonable topic before it even starts. When, for instance, the stock market crashed in October, 1987, the IBMPC administrators cut off discussion of programmed trading and its workings in several forums, including FEAR FORUM. FEAR is a discussion of the fear of computers: who fears computers; why they fear computers; and what IBM can do to reduce the fear of computers. With the widespread allegations, in the wake of the October, 1987 crash, that programmed trading contributed to and might even have precipitated the crash, FEAR FORUM was a rather natural place to talk about programmed trading and its effects on public perceptions of computers and the stock market. In the view of those who sought to discuss program trading in FEAR FORUM, such trading might easily contribute to the general fear of computers.
An October 21, 1987 append to SENSITIV FORUM by an IBMPC administrator expresses the reason for what a subsequent appender called a "pre-emtive strike":
Yes... we are "squashing" any stock market discussions for now. It seems to me that any technical discussion would certainly spill over into speculation and areas far outside the IBMPC charter.The technical side is interesting, but we are cautioning people to simply postpone that discussion until things quiet down. So jot your ideas down, refine them, and when things are quiet a discussion would be interesting.
There are several notable elements of this "pre-emptive strike". First, there are clear problems with allowing random speculation on the event without some base of facts. As a Fishkill, N.Y. appender puts in in an October 22 append, "because it's difficult to draw the line between discussing how a system works and predicting how its going to behave, and because such predictions could be a tool for *personal* gain, I think it is wise to postpone the discussion until there's a bit less serious predicting to do."
Second, the denial is temporary and clearly declared as such from the beginning. Indeed, the topic opened to discussion a week later in a new forum, PCSTOCK FORUM, which was carefully designed to ensure that discussion remained PC and IBM business related. This is fairly unusual in denial actions. Most denial actions involve clear-cut violations of the IBMPC rules. Hence topic closures are usually intended to stand. This is not always true, however, as this temporary topic closing on FEAR FORUM illustrates.
The closing of discussion concerning programmed trading in the wake of the 1987 stock market crash was, like many reviewing actions on IBMPC, controversial. While the topic was closed to discussion, the controversy was not, and the home for this meta-discussion was a meta-forum. Most meta-discussion concerning reviewing actions on IBMPC is directed to REVIEWNG FORUM. In the case of the stock market crash reviewing action, however, the controversy started with an Owego, N.Y. appenders concern about being sensitive to users and the meta-discussion started on SENSITIV FORUM:
While I fully understand the reviewer's plea to avoid discussing the reaction of the stock market(s) this week, it seems that people are being silenced when they have legitimate concern. There have been at least two appends in the last day or so which have been cut off so quickly that the authors admit they are reticent to append ANYWHERE now. I'm sure that wasn't the response the reviewers expected. It seems that we are not being SENSITIVE to the appenders who honestly want to know about the buy/sell computer *programs* which seem to have intensified the market reaction. ... I just hope that the current actions haven't put a permanent FEAR of appending into these people.
The IBMPC administrator's explanation of the reviewing action, presented above, replies to this append. Indeed, the decision to make the action a temporary one is based, at least in part, on this append's argument that there is value to such discussion. A Boca Raton appender raises (somewhat sarcastically) a larger objection almost immediately:
So from now on folks, any topic that is perfectly business related, and perfectly PC related, and meets ALL the IBMPC RULES, has the potential to be "squashed" because it MIGHT cause a couple of appends to appear that don't fit the rules. This sounds like a pretty ominous precedent is being set.
This append probably overstates the situation considerably. Prior experience told the IBMPC reviewers that this was a topic that could lead to many problematic appends. Events on the conference disk solidified that belief. In the words of a Kingston, N.Y. employee, "the fact that appends on the topic appeared on simultaneously on N fora, N on the order of 6, attacking it from non-IBMPC angles, was evidence enough that it was a chancy subject at the moment." Still, the act of restraining the topic before it had a chance to develop is legitimately debatable. In the words of the Boca Raton appender, "Give us a chance to prove we're adults here. If we blow it then purge the discussion."
Closing a topic to prevent a situation from developing may have been highly justified in this case. This is, in fact, the broad consensus of SENSITIV FORUM appenders on the subject, who favor the action by a seven to two margin (not counting the two IBMPC administrators who commented in the forum). The subsequent opening of PCSTOCK FORUM acts to further justify it as a temporary action.
The reopening re-enforces, however, the openness of IBMPC. A reviewing action taken is not necessarily a reviewing action completed. Discussion of that action on meta-forums like REVIEWNG and SENSITIV FORUM can result in changes to, and sometimes reversals of, reviewing actions. The process is described in what is, in effect, the closing append to the crash meta-discussion on SENSITIV FORUM by a Dallas, Texas appender:
A typical (major) flap on IBMPC has two parts: the "violation" and the "reviewing action" battle. The "violation", by the time the reviewers get involved, is usually well established as a problem and is not going to go away on its own.The "reviewing action" battle is more like a firecracker, made of the same feelings as the gunpowder of the "violation" but explosive because of the attempted constraint of the topic. That is, the more involved IBMPCers are in the topic, the greater will be their reaction when the topic is declared a "violation".
The final, and possibly most extreme, tool of the IBMPC reviewer is the power of management action. If a participant submits an append whose content is grossly inconsistent with existing IBMPC policies and guidelines or cannot be made to understand what it means to keep appends within the bounds of the IBMPC rules, the participants manager may be asked to handle the situation.
The role of the IBMPC reviewer is complicated by the range of enforcement powers that he or she has available. It is possible for the reviewer to delete appends, series of appends, or even entire forums. It is equally possible for the reviewer to preview every append by a given individual or to a given forum. Forums can be locked. In extreme cases, other IBM management processes can be invoked. Most often, however, the most effective reviewing action is simple expression. A problem forum thread can often be stopped or diverted with just a few well chosen words. An appender can often be convinced to change a problem submission through the simple expedient of pointing out the problem.
This door swings two ways, however. There may not be a problem where the reviewer sees one. Appenders regularly educate the IBMPC reviewer on the business relatedness of appends. Contributors to meta-forums like REVIEWNG frequently help the reviewers to refine reviewing actions and sometimes even succeed in reversing them.
Hence the IBMPC reviewing process is basically one of communication and meta-communication. The most desirable reviewing action is usually the one that the appender implements based on the persuasion of an IBMPC administrator or other IBMPC participants. Such persuasion requires an understanding of IBMPC's rules, a still better understanding of the boundary conditions of IBMPC's content, and the ability to express those rules and boundaries in a way that contributors will readily understand. There is no magic formula for being a good reviewer, but it seems clear that the individual who can do this kind of persuasion in a non-threatening way is very likely to be highly effective.