This study of IBM's IBMPC computer conferencing facility has sought to explore two broad themes:
These questions are comparative late-comers to the study. The latter question only emerged when the study of the IBMPC computer conferencing facility was largely complete when it seemed that the mechanisms that appeared to shape IBMPC might also operate in the evolution of other media. The former theme is more of a covering question for a range of other questions, including:
Three other "problems with problems" are identified in this study, including problems of incorrect assignment, balancing, and the recognition of problems in solutions. The recognition of problems in solutions is, in many ways, the most interesting problem with problems. It suggests that we frequently will not recognize the real problems with a medium until it is possible to solve them. A problem sometimes isn't a problem until, with the wisdom of 20/20 hindsight, we see what we were missing.
Balancing, by contrast, is the most difficult of these problem problems to deal with. It will be the case that the "solutions" to some problems will be more problematic than the problem it attempts to resolve. The reader is invited to ponder, for instance, how one solves 1988's biggest problem, "too much discussion", in a medium where asynchronous discussion between diversely located people is a key characteristic, application, and benefit without making the "tightness" of reviewing a bigger problem.
Incorrect assignment is the fairly common practice of assigning the problems of society on its newest media. Consider, for instance, the number of media, including plays, records, pool halls, dancing, radio, and television, the decline of social values been blamed on). The practice of incorrect assignment can be observed in such criticisms of computer conferencing as its supposed "addictiveness" (Kerr and Hiltz, 1982, and many others). There is no solution to the problem of incorrect assignment except reflection and comparisons with other media. Substantial use of computer conferencing can be considered a sign of addition or of a particularly effective means of communication. Face to face communication makes a particularly potent comparison benchmark in this and other cases.
Although a number of problems with computer conferencing are identified in this study, it is difficult to state many that are inherent to the medium. Most, like "finding things", should eventually be solvable. The truly inherent problems, including "too much discussion" and "tight reviewing", achieve this status because of inherent conflicts with other problems and solutions. It is no accident that the chapter on problems is the last chapter (other than this one) written for this study. It was the most difficult to write. For those convinced that a new medium can't solve problems without creating new ones, it will be the least satisfying. The hypothesis that successful media maximize benefits and minimize problems appears to be a correct one, however. Hence it is possible to regard the relatively small number of problems inherent to computer conferencing as a symptom of the medium's success rather than a weakness in this study.
These five questions are hardly exhaustive. They simply highlight some of the more significant elaborations of this first theme that have been addressed in this study. Other variations on the theme of exploring the nature of computer conferencing that have been addressed to some extent in this study include:
The last of these variations is, perhaps, the most important. It is the changes that IBMPC has undergone over its eight year history that led to the second theme of this study and the theory of medium as process and subsequent chapters. Computer conferencing is not a static medium. It is not static on the IBMPC computer conferencing facility, where we have observed the evolution of new features, better user interfaces, and other changes. It is not static in IBM, where the informal development of computer conferencing continues. It is not static outside IBM, where change is easy to observe on many computer conferencing facilities.
Computer conferencing does not appear to be exceptional in this regard. Although observation of computer conferencing provides the basis for the theory of medium as process presented in the early chapters of this study, the theory itself is presented in terms of other media. The observation remains, for the most part, as it was before the theoretical chapters were written (or imagined), and the job of tying the theory to the observation has so far been left to the imagination of the reader. One of the tasks of this chapter will be to make the relationship between the theory and observation explicit. It is hoped that we will, in this process, further inform the nature of computer conferencing.
We will, in the remainder of this study:
Most of the other questions which have driven this study will be reviewed, to some extent, in the process.
The assertion that computer conferencing's status relative to the traditional dichotomy of interpersonal communication and mass media is first made at the beginning of this study. It is in this first chapter that we first describe computer conferencing as a mass interpersonal medium. The evidence for this claim, initially, is fairly restricted. The primary evidence is a description of the characteristics of computer conferencing which described the medium as one which "allows large numbers of geographically distributed individuals to converse in an asynchronous manner;" which "provides for interpersonal interaction among members of a mass media audience." Secondary support is generated, however, in the observation of Reardon and Rogers (1988, p. 297) that there are media which "cannot be easily categorized as either interpersonal or mass media."
The next substantial discussion of this theme occurs as the informal typology of media. This typology, which contrasts potential recipients of a message with the mediums potential for feedback, suggests that computer conferencing is neither an interpersonal media nor a mass medium. Interpersonal and mass media can be regarded, in this informal typology, as a continuum that extends from the upper left hand corner (high potential recipients and low potential feedback, a pure mass media) to the lower right hand corner (high potential feedback and low potential recipients, a pure interpersonal media). Computer conferencing, with its high potential recipients and high potential for feedback, effectively projects a new orthogonal dimension. Whether computer conferencing is regarded as an interpersonal mass medium or a mass interpersonal medium, this typology suggests that the medium is something different.
The formal typology of media refines this view when it presents "a seventh cluster" occupying the middle ground between traditional and technological media. This seventh cluster is not visible in the cluster analysis (the computer media divide themselves among a variety of interpersonal and mass media clusters). It is possible, however, to draw a circle in the formal typology that encompasses all computer media, which occupy a common middle ground between interpersonal and mass media. This typology places computer conferencing nearer to publication media, particularly newsletters, than to interpersonal media, but the medium is clearly in the middle ground, once again a mass interpersonal medium.
A subsequent test of this typology against the perceptions of computer conference users supports and strengthens this notion of a cluster of computer media that mediate the gulf between interpersonal and mass media. Electronic mail and computer conferencing are seen, on average, as being fairly similar in this analysis, and both seem to be regarded more as interpersonal media than mass media. As viewed in the user typology, however, computer conferencing remains to be solidly in the middle ground between media. Indeed, the medium is positioned almost exactly in the center of each of the table's dimensions relative to the other media evaluated.
The analysis of the "simple query" extends this typological approach by contrasting computer mediated query with several hypothetical media. These contrasting media include slow conversation, fast correspondence, "instant" classifieds, and citizen's band "Dear Abby". Each of these media is a hybrid which stretches a conventional media in the direction of the center of the typologies as presented above. Fast correspondence would be more than interpersonal media. Slow conversation would be more like correspondence media. "Instant" classifieds would give what is clearly a mass media genre a strongly interpersonal flavor. Citizen's band "Dear Abby" combines a generally interpersonal media, citizen's band radio, with a mass media genre, advice columns.
Computer conferencing isn't quite like any of these hypothetical media, but each describes an aspect of the experience of computer mediated query and computer conferencing. One notes that this chapter is one of the first written in this study. It predates all of the subsequent typological work. Yet looking at the selection of media and the way they are modified for the comparison with computer conferencing, they almost present a capsule view of the formal typology. Two interpersonal media, a correspondence media, and two genres of publication media represent three of the four corners of the formal typologies depiction of interpersonal/mass communcation versus dynamism. All are pulled toward the middle of this representation, just as computer conferencing is.
This test of user perceptions of computer conferencing and other media is only one of several IBMPC user survey results that inform the mass versus interpersonal communication distinction on IBMPC. A second test is found in the 1988 value questions. A factor analysis of these results yields two factors. One of these factors describes the value of using IBMPC for distributing information. This factor would appear to be more mass media oriented. The second factor described a more interpersonal orientation in the use of IBMPC for various forms of discussion. Although both factors account for about 15% of the total variance for the thirteen variables used in this analysis, respondents seem to value information somewhat more than discussion.
A subsequent discussion of various genres of computer conferencing shows that information and discussion mix in a variety of ways in different characteristic processes. Computer mediated query, for instance, directs a highly constrained interaction to the business of finding information. It is rare for a thread in computer mediated query to last for more than a day. On-line user groups, by contrast, are much more strongly interactive, with themes that can last for weeks at a time. Interaction remains focused, however, around a need for information. Electronic seminars, which often mix considerable personal opinion with information, are even more interactive. At the extremes, lightening rods and polar debates frequently contain little or no information; voluntary newsletters and answers files contain only information.
The mixture of information and discussion in the characteristic processes of computer conferencing is also illustrated in the detailed discussion of SHUTTLE FORUM. Information was provided to the forum in the form of news. Discussion on the forum generally centered around analysis and meta-discussion of the forum. Information and discussion are not the only, or even the dominant, element in SHUTTLE FORUM, however. Most forum content is expressive, taking the form of mini-eulogies. Although these eulogies occur in series, they rarely refer to each other. A chain eulogy is not necessarily an interactive one. Although these eulogies generally contain detailed descriptions of events, the events are presented for their emotional symbolism more than their information content.
It is, perhaps, in the extremes of interaction on IBMPC, that the biggest problems of IBMPC, according to the 1988 survey, emerge. If discussion is a major ingredient of interaction on IBMPC, it is also the source of its biggest perceived problem. Roughly 25% of respondents to the 1988 survey cited too much interaction as one of the things they liked least about the facility. The concept of information overload, the availability of more information than one can handle, is well understood in the field of communication.
The extremes of interaction on IBMPC suggest the possibility of a new construct that can occur when interpersonal interaction is extended to a mass audience. Exploration of the same constructs endlessly by a a large audience presents the possibility of a difficult to resolve redundancy. This construct, discussion overload, suggests that people faced with more redundant discussion than they can reasonably process may have difficulty finding useful information in it.
A final form of support for the status of computer conferencing as both an interpersonal communication medium and a mass medium can be found in the discussions of rules and enforcement of rules on IBMPC. One notes with particular interest the need for meta-forums, the special forums on IBMPC that have become the focus of meta-discussion on the facility. The need for meta-forums is rooted in the dysfunctionality of meta-discussion when it occurs on forums that are primarily oriented to technical discussion.
Discussion is a good thing in these technical forums, but only when it informs the technical issues that are the focus of the forum. When discussion strays beyond those boundaries, whether it be into questions of how interaction is proceeding on the forum (meta-discussion) or debate of narrowly defined questions which may have no correct answer (polar debate), the technical discussion suffers. At its extremes, the result is murder by meta-discussion.
The problem here, once again, is a conflict between the requirements of providing information and allowing free discussion. Constraining certain kinds of discussion to meta-forums (SENSITIV FORUM and REVIEWNG FORUM, for instance) and lightening rod forums (NEWSCHAT FORUM and OS2-AIX FORUM are examples) allows the technical threads of discussion on technical forums to continue unabated while giving apparently unavoidable side discussions a home. The solution is a compromise. The result is not unlike a column or feature in a mass medium, but without the heavy editorial control generally associated with such a medium. It allows the side discussions that would simply be excluded from mass media to proceed, but not, as would be the case in interpersonal media, in line with other conversation. Meta-forums and lightening rods provide a separate conversational track for such interaction.
This mixture of interpersonal and mass medium also applies to the means by which rules are enforced. The role of the IBMPC reviewer differs from mass media roles like editor primarily in the level of interaction associated with the role. Although the reviewer has powers, not unlike those of an editor, to delete appends and forums, the primary tool of the reviewer is meta-communication via notes and meta-forums. This tool is hardly unique to the reviewer. Many participants express opinions about what is and is not appropriate content for IBMPC via notes and appends to meta-fora. The result is frequently a dialogue in which the IBMPC participants influence rules, interpretations, and reviewing actions.
The mixture of interpersonal and mass communication practice in the rules and enforcement practices associated with IBMPC reflects its mixed status. There is, as would be the case in most mass media, a central arbitrator of what is or is not appropriate content. There is, as would be the case in an interpersonal media, an informal process of enforcing rules via meta-communication among participants.
Evidence that computer conferencing has characteristics of both mass and interpersonal media is generated in virtually every major method of this study. It can be found in analysis of the characteristics of the medium. It can be found in theory based taxonomy, a numerical taxonomy of media, the taxonomy based perceptions of IBMPC participants, the structure of such strongly information oriented events as computer mediated query. It can be found in the results of the survey. It can be found in a content analysis of archival forum content. It can be found in the problems of computer conferencing. It can be found in the rules and enforcement mechanisms of the medium.
The mixture of interpersonal and mass media elements in computer conferencing is pervasive and unavoidable, and speaks to the necessity of communication theory that integrates mass and interpersonal media in a single construct. The call for such integration, as documented at the beginning of the dissertation, is already growing loud in the field of communication. It is hoped that this study provides additional impetus to this movement.
It is also hoped that the theory of media as process presented in the early chapters of this study will prove useful to such attempts at integration. It has proved useful here, where the four key words of its grammar -- mediators, characteristics, effects, and practices -- provided, after the fact, an organizing construct for a wide ranging set of observations.
Of all the chapters, the initial description of IBMPC is the most closely oriented to the description of the mediators of computer conferencing in the instance of the IBMPC computer conferencing facility. The big lesson of this chapter is that these mediators, which include mainframe computers, the VNET computer network, the TOOLS computer conferencing software, and a variety of software front ends to the facility, are not fixed entities. They are dynamically changing and often interchangeable.
VNET is only one of many computer networks. Indeed, it is but one of several within IBM. Although VNET has been in existance for over 15 years, it is only in the last five years that IBM has adopted VNET as its principle network for international communication within the company. It developed and grew, through most of its early history, through the efforts and decisions of individual IBM sites. Change in the workings of VNET has been a constant through its history, and the rise of computer conferencing has forced important changes in the way IBM maintains VNET. Dedicated network gateways that isolate network traffic away from mainframes were necessitated by increased network traffic. Continuing traffic increases have resulted in a series of upgrades to VNET's network connections. Computer conferencing traffic, in which a series of messages are broadcast to many nodes on the network, resulted in the enhancement of VNET's software with *LIST processing, which sends a single copy of a message to many destinations, breaking out copys to different nodes as near to the destination as possible. These kinds of changes continue.
The TOOLS conferencing software is but one of several computer conferencing packages within IBM. Another major package, IBM's GroupTalk computer conferencing software product (known as GRAND while still in development), is command and function compatible with TOOLS. There are differences in the style of interaction that is generally used with GroupTalk and TOOLS, but one can readily set up GroupTalk computer conferencing software to shadow TOOLS conferences. Hence the software is largely interchangeable with the TOOLS software.
The same is true, to a large extent, for the front ends used by people who follow IBMPC. FORALIST, OMNIDISK, TOOLCAT, the IBMPC reviewing software and other front ends all make IBMPC and other conferencing facilities accessible. Each has different features. Each entails a somewhat different mode of interaction with IBMPC. Some users may use more than one of these front ends depending on their needs. All, in the end, are fairly interchangeable, however. One's choice of front end is ultimately more a matter of taste and history than broad differences in what they allow you to do.
Change in both the TOOLS software and the front ends to IBMPC continue. The current form of IBMPC is becoming fairly stable and this change is increasingly more a matter of refinement than radical new features. Major changes, the more projectable of which will be discussed later in this chapter, can be expected in the future, however. Some of these changes will be driven by technology and what it makes possible. The most important ones, however, will, like subject and reference information, the preview capabilities of TOOLS, and the keyword search capabilities of the IBMPC reviewing software, probably be driven by the needs and problems of the IBMPC user community.
The characteristics of computer conferencing, by contrast, have been much more stable. IBMPC may be easier to read and append today than it was in 1981, but the essential characteristics of interaction are the same. When we define computer conferencing as a medium which "allows large numbers of geographically distributed individuals to converse in an asynchronous manner," we are describing the medium in terms of its characteristics. These characteristics do not entirely describe a medium, and they can change somewhat, but they are the essential core of what a medium is to the people that use it.
The effects of computer conferencing in the case of the IBMPC computer conferencing facility are directly addressed in two chapters: "Problems of Computer Conferencing as Experienced on IBMPC" and "Benefits of Computer Conferencing as Experienced on IBMPC". There was a time when these chapters seemed critical to this dissertation. In the end, they seem dry; mere listings of interesting effects detached from the contexts in which they matter. The more interesting descriptions of application and outcome effects of the medium are found in the chapters that describe the rules and genres of IBMPC. These chapters describe computer conferencing as a process of discovery and adaptation.
The idea of computer mediated query, for instance undoubtedly started, with the posting of a question to IBMPC that was answered with unexpected rapidity. The person who asked this first question probably needed some information and was having trouble finding it. The person or people who answered probably felt good about being able to share some bit of hard won knowledge with somebody else. The people who observed the interaction learned something unexpected, and possibly very useful. Everyone who participated in the interaction saw that the question worked, which probably made the next question easier.
The steps between that first successful CMQ event and the routine computer mediated query events of ISTHERE FORUM describe a process of genre formation. More questions were asked. Styles of asking questions started to emerge. Thank you appends were supplanted by "Thanks in advance" messages in the initial append. A convention of putting related questions into forums with the filetype "QUERY" evolved. Answers started to use reference lines and both question and answer started sharing common subject lines. CMQ forums like ISTHERE FORUM emerged. Each of these steps and many others help to define computer mediated query as a characteristic process. Each of these steps act to optimize computer mediated query as a genre which satisfies a need for quick information.
This process is inherently recursive. The need for quick information might never have been discovered without that first question and all the steps that followed. Some of the steps entailed solving problems. The "QUERY" filetype made questions and answers more obvious and accessible. The CMQ forum created a central repository for questions that didn't belong on other forums, thus reducing the tendency, on IBMPC, to have many small forums that were used, in many cases, only once. It also solved the problem of narrow questions being ignored because no one looked at the forum and the problem of off topic questions being placed on the more popular forums.
Other steps optimized the genre to the application. "Thank you in advance" may well have had the effect of reducing the number of non-information related appends in forums like ISTHERE FORUM. A more important effect, perhaps, was to encourage the person who would ultimately supply the answer. Subject and reference lines made it easier to associate answers with questions. Descriptions of the context of a problem as well as the specific question allowed people to better understand the problem and provide answers that were tuned not only to the question, but to the situation in which the answer would have to be used.
Similar solutions and optimizations can be observed in other genres. ANSWERS files make it easier to find answers to commonly asked questions. They also reduce the clutter of repeatedly asked questions on many forums. The structure of these forums, a series of appends in which a question is followed immediately by an answer, is optimized to the nature of the forum.
Voluntary newsletters provide a way for people to share information without necessarily being asked for it. The rules of these newsletters generally discourage discussion, frequently diverting it to another forum. Hence the content of E3 PROCS is discussed, along with E3 itself, in E3 FORUM and E3TECH FORUM. Hence the content of NEWSCLIP FORUM is discussed in NEWSCHAT FORUM. The process of voluntary newsletters are a matter of continuing negotiation. One such negotiation concerned the direct quotation of material in NEWSCLIP summaries. Quotes, in the wake of this discussion, are now restricted to a few lines or a paragraph, which is regarded as a reasonable "fair use" of such material.
Lightening rod forums provide a home for discussion that would otherwise distract from other technical discussion. Early failures like SOAPBOX FORUM have made the IBMPC administrators careful in their treatment of such forums. The name of the forum appears to be particularly important to such forums. The renaming of SOAPBOX to TEMPMISC radically changed that forum. Similar effects were observed when NEWSTALK became NEWSCHAT. It is common, moreover, for lightening rods to have specialized and fairly restrictive rules governing content. WISDOM FORUM, for instance, subscribes to a three complaint rule (if the forum owner receives three complaints about a subject, the subject is closed). TEMPMISC FORUM restricts the length of discussions (generally less than a week), rapidly discards older content. Discussions that appear to have enduring value are either moved to appropriate existing forums or used as the beginning of new forums.
The evolution of lightening rod forums entails more than rules, however. It also includes the style of opening an append on a particularly controversial subject with a line like "Before this discussion ends, let me say that ..." or closing it with a line like "I hope this is the last append on this subject" (a line that appears to virtually guarantee that it won't be). It is common, moreover, for appends in NEWSCHAT, for instance, to reference appends in other forums, particularly NEWSCLIP.
Meta-forums satisfy the need to discuss the process of computer conferencing on IBMPC and have evolved, over time, to meet the needs of users. This adaptation has sometimes meant opening specialized meta-forums like CONFEREN FORUM, which discusses the IBM Corporate Computer Conferencing Guidelines. Discussions on REVIEWNG FORUM frequently evolve in a highly characteristic manner. When a predictably controversial reviewing action is taken a pointer is created to REVIEWNG FORUM from the site of the editing action. This pointer has two purposes. First, it points discussion away from the forum on which the action occurred, thus allowing technical discussion on the forum to continue unimpeded. Second, it points discussion into REVIEWNG FORUM, thus providing the starting point for an event that will continue there.
Different variations of reviewing problems generally proceed through a fairly predictable set of arguments. If a bid to start a new forum is turned down or, more rarely, an existing forum is closed, it can be expected that there will be a discussion of some of the other arguably borderline forums on IBMPC. Such discussion will frequently cite some existing forum that the writer views as not being business related. If a controversial append is erased, one can expect an accusation of censorship, even if the action specifically recommended another place on IBMPC where the append could be appended without action.
One can expect, in either case, one or more responses suggesting that firmer action should be taken to make the conference more strongly business related. Someone will nearly always respond to appends from both camps with one that says, in effect, "that its IBM's ball and IBM's field. If you want to play ball you have to play by the rules." Finally, one can expect the reviewer to defend both the action and the decision not to broaden it in terms of the "boundary conditions" of IBMPC's rules.
None of this is a statement that interaction on REVIEWNG FORUM is either predictable or boring. Its generally very interesting. Still, the participants on IBMPC have varying goals, and the interaction on REVIEWNG is inevitably structured by these various goals and the optimum strategies for achieving each. The participant who wants to start a particular forum on IBMPC probably has a fairly good reason for doing so. A turn down will inevitably be taken, in part, as a judgement of the quality of the idea, even if the rejection recommends an alternative conferencing facility to which the forum would be better suited. If the appender had wanted it there he or she would have put it there first, and it is only reasonable, in such cases, to try to show how that forum is more business related than some other forum.
Other participants have similar motives. The participant who feels strongly about the value of discussion of IBMPC may have little tolerance for reviewing actions that appear to impede discussion. The invocation of words like censorship in an attempt to discourage such strategy is probably excellent rhetorical strategy under most circumstances. The participant who feels strongly about the information value of IBMPC will have a very different perspective, however, and may even encourage outright censorship in the interests of increasing the signal to noise ratio of forums that are mired in what is perceived to be trivial discussion. The reviewer, who generally has other work to do, is primarily interested in calming the discussion on REVIEWNG FORUM as rapidly as possible. The invocation of boundary conditions and the difficulty entailed in such decisions is aimed as pacifying both sides without satisfying either.
Each strategy is, then, a genre optimized approach to achieving the goals that can be expressed, and sometimes achieved, through discussions in meta-forums. The strategies for using REVIEWNG FORUM change over time. Indeed, one expects, in the wake of this study, a diminuation in the use of some strategies including invocation of boundary conditions, which will probably become somewhat less effective for a while. The strategies will not disappear (they work to well), but they will continue to change as participants learn better how to use the medium of computer conferencing and its various genres to advantage.
The content of meta-forums, including REVIEWNG FORUM and SENSITIV FORUM, is frequently aimed at problem solving. It is in these forums that the informal rules and boundary conditions of the formal rules of IBMPC are frequently worked out. These rules, and the mechanisms that are used to enforce them, make additional indirect statements about the effects of IBMPC and computer conferencing on its users. These development of these rules, like the development of genres, is also recursive. A rule that solves one problem may well cause another, and the process of rules building on IBMPC is frequently a process of balancing one problem against another real or prospective one.
The rules of IBMPC started simply enough. Indeed, the first rules of the facility were little more than a statement of the kinds of material that could be put on the facility. Only one rule, a mild reminder that participants should work together, in any way regulates interaction on the facility. This situation has clearly changed. Only a few of IBMPC's early rules survive in any form, and those that remain, like most of the rules of IBMPC, have been repeatedly modified.
This is well illustrated in the evolution of IBMPC's rules concerning meta-discussion. When meta-discussion became a problem on a series of forums on IBMPC, a potential solution was found in a ban on such meta-discussion. The solution was not perfect, but it did provide a basis in the IBMPC rules for the deletion of such discussion when it became disruptive. Such deletions hardly solved the problem, however. The occurrence of meta-discussion was a symptom of other problems that needed to be resolved. It was only through such discussion that these problems could be understood and addressed. The recurrence of meta-discussion was, then, rather understandable. A better solution was needed.
This solution was found in the creation of meta-forums, which provided a new home for such discussions, including discussion of meta-discussion and the rules that govern it. Subsequent refinements of the rule have rescinded the outright ban on meta-discussion, in part because there are forums dedicated entirely to such discussion. Today's rule simply indicates that "such meta-discussions ... will be prime candidates for removal themselves."
Another change to the rule creates an official path for resolving problems, with IBMPC participants who believe that an append should not be on IBMPC encouraged to contact the appender or the IBMPC administrator before appending the suggestion to IBMPC itself. A similar suggestion applies to the rules of IBMPC and various reviewing actions, with participants urged to contact the IBMPC reviewer before taking additional action. These modifications of the rule are intended to reduce the problem of unnecessary meta-discussion. It has been found that such concerns are frequently well understood and can frequently be resolved without starting up a forum discussion.
One notes, relative to this rule, that the meta-discussion rule makes no mention of meta-forums. This choice reflects, in part, the general preference of the IBMPC reviewers to resolve individual problems without extensive discussion. It also reflects the range of meta-forums that exist on IBMPC. Different meta-forums specialize in different kinds of meta-discussion. A delineation of these forums in the IBMPC rules would require a level of detail that is hardly required in the formal rules of IBMPC. The practice of using meta-forums for meta-discussion has the status of an informal rule on IBMPC. Such forums are used explicitly for the discussion of IBMPC. Indeed, IBMPC's administrators and forum owners frequently formally redirect meta-discussion to one or another of IBMPC's meta-forums.
This iteration extends throughout the process of using IBMPC. Subject lines, for instance, were very much the exception on IBMPC through its early years. As IBMPC volume grew, however, participants increasingly wanted to be able to skip by appends on the basis of a subject line. Discussion of this desire led to an informal consensus. An increasing number of appenders started including subject lines. An informal pressure to create subject lines emerged. This informal rule was eventually converted to a mediator when IBMPC user interface software tools like the IBMPC exec (program) started prompting for a subject line when an append was created; started reproducing the existing subject line from existing appends when a reply was made. This mediator enforced subject convention was eventually extended for append replies to include a backpointer (in a reference line) to the referenced append.
These informal rule based modifications of IBMPC's user interface software have been so successful that appends without subject lines are now the exception on IBMPC. The existence of subject and reference lines can now be assumed. Tools like FORASUMM and FORAVIEW have taken advantage of these characteristics to provide an improved interface to IBMPC. FORASUMM allows the IBMPC user to view a subject line based summary of a forum. FORALIST provides, among other features, one that will jump back to allow viewing of a referenced append with a single keystroke. Each of these features makes IBMPC easier to read. FORASUMM eases the process of skipping over lines of discussion that are of little interest to the reader. The "To Ref" feature of FORAVIEW makes it easy to give a reply context in the append it refers to.
Use of these features has probably already started to iterate back into the rules. Recent discussions on IBMPC meta-forums have suggested that IBMPC appenders need to give more thought to the content of subject lines, especially when a reply invokes a new line of discussion. One suspects that this discussion is related to the increasing dependence of many IBMPC participants on subject lines that accurately reflect the content of an append, a dependence that is ultimately rooted in tools like FORASUMM.
Neither its mediators, characteristics, effects, nor practices can give a complete picture of the nature of computer conferencing on the IBMPC computer conferencing facility. Mediators can describe the mechanisms of conferencing on IBMPC, but cannot describe the way those mechanisms are used or the reasons they exist. Characteristics can describe the essence of what computer conferencing is, on IBMPC and elsewhere, but without the detail that makes instances of any media different from other instances. Effects can describe the impact of the medium, but cannot describe the means by which that impact was achieved. Practices can describe how the mediators of computer conferencing are used to achieve various effects, but cannot describe either the effects or the mediators that made them possible.
An understanding of the nature of computer conferencing, as exemplified by IBMPC, should be found, then, in the intersection of these elements. The rules and characteristic processes of IBMPC are a product of its effects. The effects of IBMPC are a product of its characteristics and practices. The characteristics of IBMPC are a product of a set of mediators and, to a lesser extent, a set of practices. The mediators of IBMPC are selected, at least to some extent, on the basis of practices and effects.
The relationships between these elements are iterative, complex, and evolving. A problem leads to a rule may result in other problems. Those problems, when resolved with a new characteristic genre, may entail both opportunities and new problems. The opportunities may result in refinement of the genre. The problems may result in refinements to the rules. Any of these refinements may spill over into a new or modified mediator.
All of this can be observed repeatedly in this study's observation of the IBMPC computer conferencing facility. SHUTTLE FORUM never would have happened on IBMPC if the participants on IBMPC didn't regard themselves, at least in some sense, as a community. That sense of community probably wouldn't have happened if characteristic genres like computer mediated query had not been successful. The success of SHUTTLE FORUM enhances that sense of community and probably makes other genres more successful.
The magic of computer conferencing is the relative ease with which these processes can be observed. The mediators of computer conferencing are fairly easy to identify. All are inventions. The most important mediators, located in the software that drives the computer conferencing system, are readily changed, moreover. The characteristics of computer conferencing are also fairly easy to spot and surprisingly stable. There are many differences in the workings of EIES (Turoff's computer conferencing software) and TOOLS (the software that drives IBMPC), but these differences don't change the essential characteristics of computer conferencing on the two systems.
Meta-communication in computer conferencing often must be overt, and is readily observable in forum meta-discussions, overt reviewing actions, and, at least sometimes (if one asks for or is involved in it), in electronic mail operating behind the scenes. Bidding and negotiation are frequently easy to observe. Enforcement can be more difficult, especially when it is informal. Emulation can be observed only in the repeated use of a strategy. Hence while it is sometimes possible to identify exactly when a rule starts, it generally is not possible to identify the beginning of a genre.
The effects of computer conferencing can be observed indirectly in the rules and characteristic processes they engender. They can also be observed directly through the simple expedient of asking people what they are. One suspects that this relative ease of observation will continue, for computer conferencing, for as long as it remains possible to maintain a reasonable transcript of interaction.