19-Mar-82 00:34:12-PST,1421;000000000001 Mail-from: ARPANET host OFFICE-3 rcvd at 3-Apr-81 0059-PST Date: 28 Mar 1981 1320-PST From: CSD.SCHWARTZ at SU-SCORE Subject: a reference request To: farber at USC-ISI, farber at OFFICE-1, durham at CMU-10A, wulf at CMU-10A, mclure at SRI-KL, stefferud at USC-ISI, cotton at NBS-10, dcrocker at RAND-UNIX, jwalker at NBS-10, rms at MIT-AI, abrams at NBS-10, scott at SRI-KL, panko at BBNC, rudlsin at UCLA-SECURITY, ich at OFFICE-1, shamos at CMU-10A, david at UTEXAS-11, pool at BBNC, bair at SRI-KL, postel at USC-ISIF, newcomer at CMU-10A, rubenstein at SUMEX-AIM, ffm at MIT-MC, frankston at MIT-MULTICS, lauren at UCLA-SECURITY, pine at DARCOM-KA, fahlman at CMU-10A, sweer at SUMEX-AIM, geoff at DARCOM-KA, csl.bkr at SU-SCORE, csd.feigenbaum at SU-SCORE, admin.mrc at SU-SCORE, csd.schwartz at SU-SCORE Redistributed-To: stef at SRI-KA Redistributed-By: STEFFERUD at OFFICE-3 Redistributed-Date: 3 Apr 1981 Howdy At one point in the not too distant past (early 1979) each of you was involved in a brief, intense debate on the Finger program, its defaults, and computer privacy in general. I badly need a copy of that debate. Could any or all of you please send me a pointer to an on-line copy? If its non-existent, I could use a hard-copy if any of you happen to have one. Thanks very much Mike Schwartz csd.schwartz @ Su-Score ------- 19-Mar-82 00:34:12-PST,1661;000000000001 Mail-from: ARPANET host OFFICE-3 rcvd at 3-Apr-81 0059-PST Date: 28 Mar 1981 1418-PST Sender: GEOFF at SRI-CSL Subject: Re: a reference request From: the tty of Geoffrey S. Goodfellow Reply-To: Geoff at SRI-CSL To: CSD.SCHWARTZ at SU-SCORE Cc: farber at USC-ISI, farber at OFFICE-1, durham at CMU-10A, Cc: wulf at CMU-10A, mclure at SRI-UNIX, stefferud at USC-ISI, Cc: cotton at NBS-10, dcrocker at RAND-UNIX, jwalker at NBS-10, Cc: rms at MIT-AI, abrams at NBS-10, scott at SRI-KL, Cc: panko at BBNC, rudlsin at UCLA-SECURITY, ich at OFFICE-1, Cc: shamos at CMU-10A, david at UTEXAS-11, pool at BBNC, Cc: bair at SUMEX-AIM, postel at USC-ISIF, newcomer at CMU-10A, Cc: rubenstein at SUMEX-AIM, ffm at MIT-MC, Cc: frankston at MIT-MULTICS, lauren at UCLA-SECURITY, Cc: fahlman at CMU-10A, sweer at SUMEX-AIM, csl.bkr at SU-SCORE, Cc: csd.feigenbaum at SU-SCORE, admin.mrc at SU-SCORE Message-ID: <[SRI-CSL]28-Mar-81 14:18:40.GEOFF> In-Reply-To: Your message of 28 Mar 1981 1320-PST Redistributed-To: stef at SRI-KA Redistributed-By: STEFFERUD at OFFICE-3 Redistributed-Date: 3 Apr 1981 I believe that debate was part of the many MSGGROUP discussions, the archives of which you can find in on USC-ISI which you can FTP with ANONYMOUS FTP login. I just looked down there, and the relevant files are MSGGROUP#.0701-800 & MSGGROUP#.0801-0900 Now that you've got my (and likely everyone elses) curiousity sparked, why are you interested in that controversy? It was highlighted with many a flame and should be read with fire resistant line-printer or XGP paper if you intend to list it. 19-Mar-82 00:34:13-PST,3230;000000000001 Date: 5 Apr 1981 1155-PST From: CSD.SCHWARTZ at SU-SCORE Subject: reply to your note To: stefferud at OFFICE-3 Redistributed-To: DCrocker at UDEL-EE Redistributed-By: STEF at DARCOM-KA Redistributed-Date: 12 Apr 1981 Hi Stef (and anyone else who eventually gets a copy of this) Don't worry; I have no intention of publishing any of this; I am using the Finger controversy as a case study in a paper on possible effects of computer communication networks on informal scientific communication (yes, its a mouthful) for communication class I am taking at Stanford. This is not a major paper, nor is it of publishable quality. (if anyone wants to see it and how the Finger discussion was used in it, the paper is in [su-score]comm.txt ) I know nothing about AFIPS or about its topics or about your participation in them, nor did I have anything to do with the original Finger debate. I do think someone should publish a transcript of the discussion, perhaps with strong editing to cut out much of the personal garbage. However, I currently have no plans to do so. Now that I have assured you that I am NOT going to do so, It might make an interesting debate in itself to discuss such issues as ownership of that debate, the rights of the contributors, etc. Also, I feel compelled to comment on the tone of your letter. I don't want to make a big stink, but I think it was rudely phrased and that you should know that I feel that way. I originally heard of the Finger debate during a class taught at Stanford by Ed Feigenbaum, who handed out edited hard- copies of the discussion. The only reason I sent out my request was to find an unedited version so that I was certain I was representing the whole story. I was being careful to watch out for the feelings and rights of other people; if the contributors really feel that this subject and their early contributions to it are that sensitive then they should either (1) never had made the comments (2) asked that they be removed or (3) copyright them so no one else can publish them. If the debate had gone on in a public medium, such as a journal or conference session, there would be no recourse to proprietary feelings, and I don't think that there should be any difference in this case. Most importantly, though, I want you to know that there are no hard feelings; next time, though, don't assume I haven't considered these issues. You should also know that several people have asked what the heck I'm doing with this discussion; apparently you are not the only one with strong feelings about it. Finally, I have sent this response only to you because I don't have sufficient justification for cluttering other people's boxes with it. I will save it and send it on to anyone who asks for it; you should feel free to do the same. However, you should know that I intend to send a copy of your letter to me along with my response so that things don't get discussed in a vaccuum. Thanks for your interest. Its a good sign, I think, of professional concern with these which will be plauging computer science and society until they are thoroughly discussed. Sincerely, ------- 19-Mar-82 00:34:13-PST,1191;000000000001 Mail-from: ARPANET host OFFICE-3 rcvd at 14-Apr-81 0018-PST Date: 7 Apr 1981 2050-PST Sender: STEFFERUD at OFFICE-3 Subject: Re: reply to your note From: STEFFERUD at OFFICE-3 To: CSD.SCHWARTZ at SU-SCORE Message-ID: <[OFFICE-3] 7-Apr-81 20:50:41.STEFFERUD> In-Reply-To: Your message of 5 Apr 1981 1155-PST Fcc: SAVED.MESSAGES Redistributed-To: stef at SRI-KA Redistributed-By: STEFFERUD at OFFICE-3 Redistributed-Date: 14 Apr 1981 Hi - Your comments are very well taken, though you may have over reaacted to my message as much as I to yours. My reaction was predicated more on what other strange folks have wanted to try, than on anything you have said or done. I plan to print your report and perhaps use it in my UC Irvine class where I hope to turn someone on to doing the sort of thing you have done. I would write more but at the moment I am connected through DATAPAC/TELENET to ECL and through ARPANET (TELNET) to OF3 to answer this. Awkward at best, and I need to retire for the night so I can do a good job on my Presentation tomorrow. I am at th IFIP Int'l Symposium on Computer Message Systems in Ottawa Canada.. Cheers - Stef 19-Mar-82 00:34:13-PST,466;000000000001 Date: 8 Apr 1981 1319-PST From: CSD.SCHWARTZ at SU-SCORE Subject: Re: reply to your note To: STEFFERUD at OFFICE-3 In-Reply-To: Your message of 7-Apr-81 2050-PST Redistributed-To: DCrocker at UDEL-EE Redistributed-By: STEF at DARCOM-KA Redistributed-Date: 12 Apr 1981 hi stef -- thanks for the reply. If you do end up using the report, please let me know how it is received. Enjoy IFIP (whatever it is). Good to talk with you. mike ------- 19-Mar-82 00:34:13-PST,1265;000000000001 Mail-From: STEF Received-Date: 12-Apr-81 1641-PST Date: 12 Apr 1981 1640-PST Sender: STEF at DARCOM-KA Subject: Re: reply to your note From: STEF at DARCOM-KA To: CSD.SCHWARTZ at SU-SCORE Cc: STEFFERUD at OFFICE-3 Message-ID: <[DARCOM-KA]12-Apr-81 16:40:59.STEF> In-Reply-To: Your message of 8 Apr 1981 1319-PST Fcc: SAVED.MESSAGES;1 Hi Mike - I have FTPed your file and have looked through it with EMACS, to remove any garbage that would mess up a print run, and to scan its content. I have taken the liberty of changing Marty Turoff to be Murray Turoff in all instances, and I have tried to insert a few missing footnote citations where they seem to belong. I have not changed anything of any consequence. You can find the file in [KA]csd.schwartz if you want to check over what I did to it. All the footnote citations are now in []. So far, without careful reading, I am rather impressed with it and I plan to print it and use it as an example for my class. I will let you know how it goes. I would like to be able to get some UCI students to do something like his, but I am not sure how to get the transcripts to them, other than by tape reel, since UCI is not on ARPANET. I will keep you posted. Stef 19-Mar-82 00:34:13-PST,46854;000000000001 Mail-From: STEF Received-Date: 12-Apr-81 2042-PST Date: 12 Apr 1981 2042-PST Sender: STEF at DARCOM-KA Subject: [STEF at DARCOM-KA: Re: reply to your note] From: STEF at DARCOM-KA To: DCrocker at UDEL-EE Message-ID: <[DARCOM-KA]12-Apr-81 20:42:52.STEF> Fcc: SAVED.MESSAGES;1 Hi Dave - Here is my reply and a copy of the paper FYI. Stef Begin forwarded message - - - - - - - - - - - - - - - - - - - - 1265 chars Subject: Re: reply to your note Date: 12 Apr 1981 1640-PST From: STEF at DARCOM-KA - - Sender: STEF at DARCOM-KA To: CSD.SCHWARTZ at SU-SCORE Cc: STEFFERUD at OFFICE-3 In-Reply-To: Your message of 8 Apr 1981 1319-PST Hi Mike - I have FTPed your file and have looked through it with EMACS, to remove any garbage that would mess up a print run, and to scan its content. I have taken the liberty of changing Marty Turoff to be Murray Turoff in all instances, and I have tried to insert a few missing footnote citations where they seem to belong. I have not changed anything of any consequence. You can find the file in [KA]csd.schwartz if you want to check over what I did to it. All the footnote citations are now in []. So far, without careful reading, I am rather impressed with it and I plan to print it and use it as an example for my class. I will let you know how it goes. I would like to be able to get some UCI students to do something like his, but I am not sure how to get the transcripts to them, other than by tape reel, since UCI is not on ARPANET. I will keep you posted. Stef *END MSG Message-ID: <[DARCOM-KA]12-Apr-81 16:40:59.STEF> End forwarded message (* Note that this file was originally in BRAVO format. I have (* cleaned it up a little bit, but dont be surprised to find garbage (* in the text. -- mjs. Informal Scientific Interaction and Computer Communication Networks Introduction. During the last twenty years, researchers of scientific and technical communication have concentrated on information flow and personal interaction. Earlier research examined the effects of each particular medium, while more recent research emphasizes the informal, interpersonal scientific communication channels and networks.[1] A seminal writer of the new branch of communication research, Derek Price, re-invented the term "invisible college" to describe one such structure.[2] Other researchers[3] have continued the formal investigation into informal communications structures. Also during this period, a new medium for communication appeared: the computer. The new medium was quickly adopted as a subject of research by several groups who examined the impact of the new medium. So far, though, no one has carefully examined the effects of computer-based communication and computer networks on the structure and social organization of human communication and human networks. There have been a few speculative articles[4] but little empirical examination has been performed. Unfortunately, the scope of this project does not allow for such examination. Rather, I will speculate, attempting to synthesize information and network theory with what is known about the computer as a communication medium. I will generate several hypotheses about the social organization of, and communication within, science as it shapes and is shaped by the computer as a medium for communication. Some Preliminaries Before beginning our discussion I need to introduce a few concepts and terms which will make the mechanics of the paper much easier. (a) Computer-based communication can be categorized as "in-machine" communication or as "network" communication. In-machine communication refers to interactions between two or more users of a single machine, for example through a message-drop capability. Network communication implies interaction between users of different machines which are electronicly linked. Modern computer message systems make these two cases nearly transparent to the user. Also, in the context of our discussion, the distinction between the two is not at all significant and I will therefore treat the two cases identically. I mention this here because (1) the distinction is very important for some issues which are not covered in this paper but with which readers of this paper may have concern, and (2) I want to avoid confusion if a reader's pre-conception of computer-based communications includes only one of the two categories. (b) The term "Computer Conferencing" (CC) is employed extensively in the literature, particularly by Hiltz and Turoff.[5] Conferencing, however, implies a restricted genre of communication. Therefore, I employ the term and the abbreviation to mean not only conferencing but also the general notion of computer-based communication. (c) The majority of my personal experience, much of my personal bias, and most of my case studies are based on the store-and-forward packet-switched computer network[6] known as the ARPANET. "The ARPANET is an operational, resource sharing inter-computer network linking a wide variety of computers at Defense Advanced Research Projects Agency (DARPA) sponsored research centers. . . . The ARPANET originated as a purely experimental network in late 1969 under a research and development program sponsored by DARPA to advance the state-of-the-art in computer inter-netting. The network was designated to provide efficient communications between heterogeneous computers so that hardware, software, and data resources could be conveniently and economically shared by a wide community of users. As the network successfully attained its initial design goals, additional users were authorized access to the network. Today the ARPANET provides support for a large number of DoD (Department of Defense) projects and other non-DoD government projects with an operational network of many nodes and host computers. "Following the successful accomplishment of initial ARPANET design goals and the expansion of the network, it was considered appropriate to transfer the responsibility for operation of the ARPANET from DARPA to the Defense Communications Agency (DCA). In July 1975, the DCA became the operational manager of the ARPANET.[7] Appendix C contains a map of the ARPANET. For technical details, please see the documents authored by Karp, Roberts, the IEEE, and the Defense Communications Agency listed in the bibliography Distinguishing the Medium From the Message. Most interactions affect and are affected by the medium and the social structure in which they occur. The same words, presented once orally and once in written form, can communicate significantly different messages. Also, a telephone call or chance meeting between two previously unintroduced experts may alter the human communication networks in which they and their colleagues operate. These affects can be roughly divided into two categories: those related to the physical medium itself and those related to the structure within which a medium functions. The division between these categories is not clear cut or well-defined, but in this paper I am concerned primarily with structural issues. But before we discuss those, we must understand which affects belong in each category. To clarify this distinction, imagine a one-to-one replacement of all communication carried out on one medium, for example the telephone, with a computer. Imagine further that no structural changes occur. In the telephone-replaced-by-computer case, this would mean that computers are still answered by switchboards or computer answering machines, there is no storage facility, and communication must be synchronous. Under this scenario changes in interactions will appear which are attributable to the changing medium. The non-structural changes associated with computers, have been thoroughly examined in the communication literature and are briefly scanned below. Computer-based communication widens some channels and narrows others. This phenomenon accounts for most of the consternation new users feel when first interacting with other people through this medium. "What happens to new users, without their conscious recognition in most cases, is that they are faced with a kind of "culture shock" in which all of the very complex 'rules' for combining the various kinds of communication channels . . . do not work, because the nonverbal channels are missing, there are some new channels or means of communication available, and the rules or possibilities for using the written equivalent of the spoken verbal channel work differently."[8] For example, body language and turn-taking cues, both of which are important in face-to-face or telephone conversations, are removed. Pauses and voice inflections are also eliminated, often creating uncertainty. While researching this paper, a stranger on the ARPANET referred me to Jake Feinler at the ARPANET Network Information Center (NIC) located not far from my home. I communicated with Jake Feinler several times using computer-based mailing systems, during which we agreed to meet for lunch. On the meeting date, I telephoned this person to confirm our appointment and discovered, much to my surprise, that Jake Feinler is a female. Whether or not this cue should have any effect on our interactions is a separate issue. In this case the missing cues were particularly blatant, but the jolt I experienced reinforces the notion that absence of cues and psychological tools employed in traditional communication channels can have dramatic effects on individuals' communication. The computer medium can be highly impersonal. Reinforcement and feedback during traditional communication are often generated nonverbally, and computer conversation restricts those channels. On the other hand, experience with the EIES computer communication system implemented by Murray Turoff[9] and with the ARPANET indicate that conversations need not be cold.[10] Often a person will add a personal postscript to a sentence or paragraph, or schedule more time at a conference to meet, in person, an individual with whom she has only communicated with via computer. If computer conversations can be held anonymously than a large set of previously unavailable channels are opened. (EIES includes anonymous capabilities; most ARPANET sites do not.) A junior member can suggest a radical idea or contradict a senior member's idea without being ignored or chastized. A senior member can propose an alternative without losing face should the proposal fail. The narrowing of channels often requires more effort to transmit the same amount of information. A hand-waving or chalkboard-scibbling explanation is often more expedient than a typed conversation. On the other hand, the relative slowness of typing encourages more carefully considered replies. In general, though, it may not be useful to judge one medium by the criteria normally applied to another. CC may not be a surrogate for any other medium, and therefore its characteristics should not be expected to mimic other media. At the same time, though, it should not be evaluated only on its own terms.[11] Information, Communication, and Information Resources. If we want to discuss the effect of CC, then we should distinguish among information, communication, and information resources. EIES has only the capability for communication. The only "information" it contains is transcripts and records of communication activities. On computer science research machines, such as many of the ARPANET, the distinctions become very murky. The integration of users, files systems, distance-indifferent communications links, mailing systems, and the ability to connect one's terminal to a machine operating across the country make distinctions between information and communication almost impossible. Fortunately, the distinctions are not critical as far as we are concerned. The significant point is that computer-based communication must be broadly defined if it is to be meaningful. Of course, there are some behaviors which we can quite certainly label communication. Also there are features of computer facilities which we could selectively add or remove which will affect communication activities. In fact, until quite recently networking capabilities were added to computers as an afterthought rather than as a fundamental part of the machine. I will sometimes speak, then, as if communication is a well-defined and separable activity that occurs in a computer context. Throughout this paper, though, be warned that communication is not always or even often separate from other activities. Case Studies. As a common ground upon which to base further discussions, this section presents four examples of CC, each of which raises slightly different issues. All of these case occurred on the ARPANET. The scope of this project does not allow me to investigate cases which have arisen in other networks. This has good and bad effects on the discussion. On the negative side, there is a lack of generality to the discussion The ARPANET has certain physical characteristics, such as free connect time, which other nets do not have, and it lacks features, such as anonymous mail capability, which other systems do have. Also, the primary user group of the ARPANET is sophisticated computer scientists, a relatively homogeneous and not terribly representative sample. On the positive side, though, an expert user group reduces the discussion to more managable proportions by eliminating the factors associated with computer unfamiliarity. Also, most ARPANET sites evolved from computer facilities with no network capabilities into facilities participating in the network. While this is not an adequate experimental control, it does allow comparisons. THE FINGER CONTROVERSY. The first case, known as the Finger controversy requires an explanation of the program known as Finger, some version of which exists at almost all ARPANET-connected computer facilities. By executing the program Finger and providing a user indentification name, computer users can receive information about other computer users. The information most often provided is (1) whether the user is currently using, or "logged into," the system, (2) if not when the user stopped using, or "logged off," of the system, (3) whether the user has any mail in her mailbox, (4) if there is not mail, when the user last read mail, and (5) if there is mail, the most recent sender. This information can be quite useful in locating a colleague or in confirming the receipt of a letter. However, some members of some computer facilities considered the availability of such information an invasion of privacy. To solve this problem, the designer of the Carnegie-Mellon University (CMU) version of Finger created two variables which each user could change for himself. One variable allowed or disallowed the reporting of computer use information and the other did the same for mail information. The initial element of controversy was the default, or initial setting of these variables, or "switches." The Finger developer opted for the most private default, but some CMU computer scientists immediately complained that this decision unnecessarily closed off one more channel of communication. The internal debate at CMU was intense enough to catch the interest of the MsgGroup, an abbreviation for "Message Group," an organization of people who share the interest of discussing controversial computer-related topics. They communicate with each other by sending copies of the same letter to every person in a common address file. The messages are also archived in one location so that multiple copies need not be stored for long periods of time. As might be imagined, the topics of privacy, openness of research facilities, and computer paranoia generated heated discussion. Within eight days approximately eighteen people contributed more than fifty pages of debate. No conclusions were reached but the debate is fascinating to read. (Appendix A contains the transcript.[12]) I think that it would be fair to speculate that a debate of the same depth of analysis and breadth of participation, if undertaken without computer-mediated communication, would have taken months of letter writing or thousands of dollars in plane fares and conference expenses. THE EDITOR DISCUSSION. The Finger discussion was carried on in a group that already existed, but discussion need not always develop in pre-existing groups. In early April of 1980, J. Q. Johnson of Stanford University created a group to discuss and collect information about computer text editors. He started the group by mailing to a set of people whom he expected to be interested and asked them to forward his message to anyone they thought might be interested. I asked that my name be added to the distribution list, even though I was not expert enough to contribute to the discussion. As will be discussed later, this ability to look over the shoulders of an on-going discussion can be very informative. Johnson acted as the manager, clearing house, and archive location for the discussion.[13] In this case we can see that even in CC it is often necessary or desirable for a single person to manage or lead a discussion. During an eighty-four day period approximately forty people contributed more than 120 messages to the discussion. (Appendix B contains the transcript.) Note specifically that the existence of the discussion spread by word of mouth and that it was observed by at least one and I would guess many persons who did not actively contribute. REQUESTS FOR COMMENTS.[14] The bureaucracy surrounding the ARPANET includes the ARPANET Network Information Center (NIC). The NIC collects information and distributes network documents for use by network officials and users.[15] The NIC shares a computer with the Stanford Research Institute in Menlo Park, California,[16] and thus is itself part of the ARPANET community. When a proposal for a technical change to the net is being considered, or when an undeveloped idea for the network is conceived, the maintainers and developers of the network like feedback from the user community. The NIC has implemented a computer-based structure for generating such feedback. The structure is entitled Request for Comments (RFC). When the idea or proposal is ready for public perusal, a text file is placed in a location of the NIC's computer, indexed in a subject file, and then publicized. The publicity also requests any comments which readers of the proposal may wish to make and specifies a person to whom responses should be sent. The access system at the NIC's computer is such that anyone who has access to the ARPANET also has access to the RFCs. Of course, not everyone who receives notification will read an RFC, and not everyone who reads an RFC will have anything to say about it, but those who do read and who do wish to contribute have a quick, painless way to do so. At the same time the proposer or thinker can essentially put his ideas on the table in an open forum and receive the input of any interested contributors. It is not uncommon for an RFC to go through several request-comment cycles, particularly if the proposal is an important one.[17] THIS PAPER ITSELF. The research behind and the preparation of this paper have employed the ARPANET and a local Stanford Computer Science Department network. When originally investigating a topic, I used the ARPANET to send an initial query to electronic bulletin boards across the country, asking for references to the history of the ARPANET. I received approximately twenty replies, half of which provided sources and half of which requested pointers to the on-line version of this paper when it was completed. When I realized that my paper copy of the Finger controversy was incomplete, I sent messages to approximately twenty people who had been involved in the debate. Within hours I had several offers of old paper copies, one pointer to the computer-accessible archive, and a handful of slightly worried inquiries as to why I was interested in bringing up the debate again. The interactions I had with others over the network was not crucial to the development of the paper, but they certainly helped. They also contributed to my personal understanding of how computer communication is different than other conventional interactions. Scientific Communications and Social Organizations. Social organization can determine communications systems and communication systems can determine social organization. For example, the National Institute of Health has funded the SUMEX-AIM computer and placed it on the ARPANET in part to build "a collaborative research community. The aim is to encourage individual participants not only to investigate applications of artificial intelligence in health science, but also to share their programs and discuss their ideas with other researchers."[18] This belief is a keystone of the work of Hiltz and Turoff. "Geographically dispersed networks of scientists working in the same specialty area can be viewed as the prototypical 'production organism' of science, in which the 'product' is scientific knowledge and the social organization depends almost entirely on the communication system."[19] The network flow theory of scientific communication defines groups in terms of their communication behavior. Clearly, then, the creation or addition of a new medium of communication creates new possible groupings and communication behaviors. Computer communications may affect current communication structures, such as invisible colleges or laboratories, or it may allow new structures, or it may make communication organizations and behavior so flexible as to antiquate the notion of structures in communication. In actuality, all three of these possibilities have occurred in different contexts and to different degrees. A survey of the history of the theory of scientific communication structures shows that they have strengths and weaknesses when faced with the advent of CC. They are weak in that, as they have so far been presented, they are too rigid to explain or incorporate some to the results of computer-based communication. On the other hand, the fundamental concepts, with the rigid restrictions removed, are very powerful and go a long way towards helping us understand or imagine the impact of computer-mediated communication. In fact, one can view the history of scientific communication structures theory as successive stages in a movement away from rigid structures and toward flexible concepts. The advent of scientific CC merely requires one more level of abstraction from an otherwise solid conceptual framework. In the last two dozen or so years, increasing emphasis has been placed on the informal communication organization of science. In 1963 Derek Price published Little Science, Big Science, a seminal contribution to the theory of informal scientific communication. Price argued that contemporary formal structures are insufficient to (1) protect intellectual property, (2) confer prestige and rewards, and (3) communicate effectively. Price then logically deduced from rough figures and intuition a likely structure and size of a communication organization that might evolve to meet these needs. He re-invented the term "invisible college" to refer to this organization of about one hundred respected scientists who found ways to interact through conferences, summer schools, preprints, and pre-preprints. These media could be effectively employed, he argued, by roughly one hundred people. Simultaneously, these sub-specialists were creating this medium to fulfill their psychological and technical needs. For several years Price and other communication researchers searched for incarnations of hypothetical invisible colleges. They had mixed success. They discovered networks of individuals which appeared to meet Price's criterion, but the organizations were not as rigid as Price has implied. In 1972, Diana Crane published Invisible Colleges, which described two informal organizations which which could be labelled invisible colleges. In her book, though, Crane significantly de-emphasized the prestige role of Price's original hypothesis. I believe she did so for the following reason. The conferral of prestige and rewards is not strictly a function of informal scientific communication. Communication behavior is certainly involved, but the process is slower and more formal than even organized communication. Furthermore, the conferral of prestige can be well accomplished by the traditional journal and conference media. A six month delay in adulation is less troubling to the scientific community than a six month delay in communication of an important result, particularly if the intellectual property right is not being challenged. If the prestige function of an invisible college is removed, then the other two functions of invisible colleges, property right protection and technical communication, can be reduced to one function, for if the communication mechanism is effective enough to communicate results promptly, then that system will be effective in communicating the discover's right to that important result. Delays create conflicts in property rights; if there are fewer delays there will be fewer conflicts. Price's invisible college concept was generalized by Crane to the broader category of information and communication organizations rather than organizations to meet more specific needs, such as prestige conferral. David Lingwood, in a Stanford University Communication Department Ph.D. thesis, briefly summarized additional attempts to locate invisible colleges. He concluded that while there was supportive evidence for the existence of invisible colleges, there was no strong evidence that they were of the size predicted by Price. Lingwood, therefore, set out to find informal scientific communications of any size. He reasoned that the choice of size was arbitrary, since the size of a potential college was determined by the measure of connectivity. Simultaneous attendance at a conference could be considered interaction, as could citation in a paper. The size of the college found using these two measures would clearly be different, but the the structures are not different; our interpretations of them are. Lingwood also points out that "The search for the 'perfect' invisible college loses importance when we consider the uses to which knowledge of interpersonal communication systems among scientists might be put."[20] Lingwood empirically and statistically looked for whatever organizations he might find. In his study, he discovers several groups of approximately ten people each. He tries to decide whether (1) these are invisible colleges in themselves, (2) these are portions of larger invisible colleges, or (3) invisible colleges exist at all. He finally concludes that no such conclusion can be drawn. More recently communication thinkers have realized that "The membership and the problems of various research networks tend to overlap in a complex manner. And the personnel involved in any one network tend to change rapidly, often being drawn from a number of different specialties and disciplines."[21] Crane hinted at this when she explained that "Science as a whole appears to consist of hundreds of research areas that are constantly being formed and progressing through [Price's] stages of growth before tapering off."[22] Even here, though, Crane attributes too much regularity and consistency to communication organizations. The emphasis, I feel, should be on the existence of many levels and many scales of informal interaction among scientists rather than focussing on or searching for a specific structure. Computers and Scientific Communication. The introduction of informal computer-based communication structures makes this shift in emphasis even more crucial. Some CC systems are intentionally structured on the principles of human communication behavior.[23] Other CC systems, such as the ARPANET, have been implemented with little or no structure in mind. However, capabilities for users to create structure do exist. The Finger controversy and the Editor discussion are examples of this. Whether or not CC systems intentionally affect communication patterns, certain effects will result from the introduction of the new medium. This section of the paper draws on what little is known about the effects and add a healthy dose of educated guesswork in an attempt to focus further discussion and research on what I view as the most important issues. The first change is that CC limits the significance of the size of the group involved in any communication activity. The Editor discussion, for example, involved forty people, but it could just as easily involved half or twice as many people. The distinctions between one hundred member invisible colleges, three hundred person discussion groups, or the dozen or so people who requested copies of this paper are far less significant if the interactions are computerized than if they are carried out in any other medium. Furthermore, distinctions between on-site and long distance interactions becomes less important. During the late 1970s, David Wilkins, then a graduate student in Stanford University's Computer Science Department, developed a thesis topic for which no advisor was available at Stanford. However, the leading expert in the field was a professor of Computer Science at Carnegie-Mellon University. The two were able to establish an advisee-advisor relationship using the ARPANET and to produce an acceptable thesis.[24] In some sense, the interactions between these two could be viewed as a small version of Price's or Crane's invisible college, a version that probably could not exist without computers. Computer networks relax constraints on interactions that other media can not relax, including interaction as a function of group size and geographic location. Scientific specialties are usually distributed geographically, for each researching institution wants a member of each specialty on its staff.[25] Price explained that as part of the invisible college mechanism, "ways and means are being found for physical juxtaposition of the members. They seem to have mastered the art of attracting invitations from centers where they can work along with several members of the group for a short time."[26] The traditional means for bringing researchers together physically are either expensive, infrequent, and time consuming, as with in-person meetings, or brief, capricious,and shallow as with telephone or mail correspondence. CC adds more possibilities by allowing long-term, in-depth interactions without the need to be in the same geographic location. The Request for Comments procedure is one example where this advantage was put to effective use. CC is cheap[27] and is not restricted by conference time schedules or the end of vacation. On the other hand, face-to-face interactions of a conference and the immediate feedback form a telephone call are both valuable and are both missing from computer-based communication. Another feature of CC is what I shall call the looking-over-the-shoulder phenomenon. My "participation" in the Editor discussion is a practical example. I had my mailbox address added to the distribution list and from then on I "watched" as the "conference" went on. I made no entries into the general discussion, although I did send some private comments and asked questions of several individuals involved in the group discussion. If an organization had held a physical conference on the same topic, it would not only have been more difficult for me to attend but also would have placed an added burden on the conference with no added benefit. Since the debate was machine-based, though, I could benefit from the discussion without requiring an added seat, an extra program, or a plane ticket. CC is far more flexible than other media for informal scientific communication. A discussion can be carried out among two or two hundred people with essentially no added difficulty. It could also last two weeks or two years. Members can "leave" or "join" a discussion by merely editing a distribution list, which is very important if human networks are constantly shifting. This flexibility also assists in establishing new formal or semi-formal organizations. The expense and overhead required to start even a mimeographed mini-journal or to organize a first conference can deter a necessary invisible college from forming. With CC the group can carry out most of its functions until it is more certain that formal structures are necessary and worth the effort. I'm sure that many struggling journals would much prefer to be computerized than to be printed. Computer-based communication can allow invisible college-level interaction on topics which might not otherwise be worth the effort; Requests for Comments are an example. CC can also affect even the most formal communication organizations in such tasks as interactive or co-operative refereeing of papers and grant proposals. Unfortunately, the role of CC in formal communication must be left for another paper. Computers also affect the formality of communication. After users have gained experience on the system, they become less concerned with well-punctuated sentences or correct spelling in their communication. They also become increasingly less inhibited about sharing a rough draft or even an abstract of a paper.[28] I suspect that this is due to the cultural prejudice that nothing is final or even significant until it is on paper in some formal presentation. This creates its own set of problems, however, for a critic or commentor may have difficulty reviewing an on-line paper or book if that version is subject to constant modification. The printing press and formal communication channels still must provide definitive versions. Up to this date, the ARPANET has been relatively informal. That is, there has been little attempt to regulate the content of network traffic even though the net is government property and funded only for research. My use of the net, for example, in preparing this paper is more than likely a technical breach of policy, but this kind of thing has gone on since the day the net started. The ARPANET hierarchy is discussing the possibility of a more formal, rigidly enforced policy. I feel that formalization is a bad idea. CC inherently limits intimacy and personability in so many ways that all forms of informality and personability should, if not actively encouraged, be allowed. Joshua Lederberg, Nobel Prize winner and current president of Rockefeller University, was active in the Artificial Intelligence in Medicine research group, the primary users of the SUMEX-AIM computer on the ARPANET. He made the following statement regarding intimacy on informality on the ARPANET. "This intimacy can and should be supported by encouraging the occasional use of the [CC] system for arranging personal rendezvous. The trivial costs of such diversions are more than compensated by the enhanced efficiency of a worker who becomes adept at the use of [CCs] as if they were an extension of his own voice and handwriting."[29] The informality of computer network communication can have interesting results, as some of my experiences while writing this paper may show. In the course of my preliminary research I came across articles written by Dr. Douglas Engelbart, an active and highly respected researcher in CC, and by Dr. Joshua Lederberg, who I have already mentioned. These articles were very helpful, but they were quite old and located in obscure works. I wanted to find any other relevant works by these two authors, particularly more recent ones, and I was concerned that I might not find them if they did exist. I sent informal notes to each, requesting references that they might have to their own work. Both replied quite promptly and quite helpfully. I later realized that I was researching a paper which I did not expect to be an important scientific contribution and yet I was requesting the assistance of two very eminent and very busy people. Through any other medium I am certain that I would not have had the audacity to make such a request, but the dual factors of ease of communication and the informality of the medium made such a request nothing more than a routine inquiry. Some Closing Thoughts. Informal scientific communication based on computers will be a significant new factor in the operation and study of scientific communication. Costs of computers have dropped by an order of magnitude every five years for the last three decades and they have a few decades left to go. Hiltz and Turoff estimate that by the 1990s one out of three scientists will have access to computers with general computational and communications ability.[30] I suspect that even this estimate is conservative. Both the disadvantages and advantages of computer-mediated communication are going to be immense, and much work needs to be done now to provide the knowledge needed to effectively exploit this new medium. Computer networks are not yet in wide use. This provides us the opportunity to examine aspects of computer-based informal communication now, before customs or economic factors become unalterable. The limited number of CC systems in operation provide an excellent medium for the study of informal communication behavior, for the transcript of the discussion is an exact replica of the interaction, the type of replica which Price, Crane, or Lingwood needed. Of course, the computer is then interfering with the study, but with the acknowledgment of its weaknesses as a research tool, computers not only provide us with a new tool to communicate but also with a new tool for the study of informal scientific communication. N O T E S 1. David A. Lingwood, Interpersonal Communication, Research Productivity, and Invisible Colleges. A Dissertation at Stanford University Department of Communication, pp. 7-9. 2. Derek J. de Solla Price, Little Science, Big Science. 3. Other researches include Diana Crane, David Lingwood, and N.C. Mullins. See their titles in the bibliography. 4. Joshua Lederberg, "Digital Communication and the Conduct of Science: The New Literacy." Proceedings of the IEEE, 66:11 November 1978. 5. Starr Roxanne Hiltz and Murray Turoff, The Network Nation. 6. See Lawrence G. Roberts, The Evolution of Packet Switching," and the IEEE, Special Issue on Packet Communications Networks, Proceedings of the IEEE, 66:11 November 1978. 7. Defense Communications Agency, Arpanet Information Brochure, p. 1. 8. Starr Roxanne Hiltz and Murray Turoff, Op. Cit., p. 81. 9. Ibid. 10. Starr Roxanne Hiltz, "The Human Element in Computerized Conferencing Systems." Computer Networks, p. 423. 11. Peter Lykos, (ed) Computer Networking and Chemistry, p. 56. 12. Accessible over the ARPANET. The relevant files are [usc - isi] MsgGroup#.0701-0800 and [usc-isi] MsgGroup#.0801-0900. Use the Anonymous Ftp login convention. 13. Accessible over the ARPANET. The relevant files are [su -score] editormail.txt. 14. Defense Communications Agency, Arpanet Information Brochure, p. 25. 15. Ibid., p.26. 16. By this date they may be own their own machine. 17. Personal interviews with Jake Feinler, director of the NIC, and Vic White of SRI and Stanford University. 18. Peter Lykos, Ibid., p. 192-3. 19. Starr Roxanne Hiltz and Murray Turoff, Op. Cit., p. 19. 20. David A. Lingwood, Op. Cit., p. 177. 21. Ina Speigel-Rosing and Derek de Solla Price, (eds) Science, Technology, and Society: a Cross-Disciplinary Approach, p. 110. 22. Diana Crane, Invisible Colleges: Diffusion of Knowledge in Scientific Communities, p. 12. 23. Starr Roxanne Hiltz and Murray Turoff, Op. Cit. 24. Personal communication from David Wilkins. 25. Ina Speigel-Rosing and Derek de Solla Price, Loc. Cit. 26. Derek J. de Solla Price, Op. Cit., p. ?? 27. Murray Turoff and Starr Roxanne Hiltz, "Meeting through your Computer." IEEE Spectrum, May 1977, p. 60. 28. Ibid. 29. Joshua Lederberg, Op. Cit., p. 1316. 30. Starr Roxanne Hiltz and Murray Turoff, Op. Cit., p. 247. B I B L I O G R A P H Y Blanc, Robert P. and Cotton, Ira W., (eds) Computer Networking (New York: IEEE Press), 1976. Carhart, Raymond E. et al, "Networking and a Collaborative Research Community: A Case Study Using the DENDRAL Program," in Lykos, Peter, (ed) Computer Networking and Chemistry (Washington, D.C., American Chemical Society), 1975. Cole, Jonathan R. and Cole, Stephen, Social Stratification in Science (Chicago and London: University of Chicago Press), 1973. Conrath, David W. and Bair, J. H., "The Computer as an Interpersonal Communication Device: A Study of Augmentation Technology and its Apparent Impact on Organizational Communication." Proceedings, Second International Conference on Computer Communications, August 1974, 121-128. Crane, Diana, Invisible Colleges: Diffusion of Knowledge in Scientific Communities (Chicago and London: University of Chicago Press), 1972. Defense Communications Agency, Arpanet Information Brochure, Washington, D.C., February 1979, AD-A066 226. Engelbart, Douglas C.; Watson, R. W.; and Norton, J. C., "The Augmented Knowledge Workshop". Proceedings AFIPS National Computer Conference, June 1973. Also reprinted in Blanc, Robert P. and Cotton, Ira W., (eds) Computer Networking (New York: IEEE Press), 1976. Fano, R. M., "On the Social Role of Computer Communications." Proceedings of the IEEE, November 1972. Also reprinted in Blanc, Robert P. and Cotton, Ira W., (eds) Computer Networking (New York: IEEE Press), 1976. Farber, David J., "Networks: An Introduction." Datamation 18:4 April 1972, 36 - 39. Hiltz, Starr Roxanne, "The Computer Conference." Journal of Communication, Summer 1978, 157-163. ----------, "The Human Element in Computerized Conferencing Systems." Computer Networks 2:6 1978, 421-428. ---------- and Turoff, Murray, The Network Nation (United States: Addison-Wesley Publishing Company), 1978. IEEE, Special Issue on Packet Communications Networks, Proceedings of the IEEE, 66:11 November 1978. Johansen, Robert; Vallee, Jacques; and Sprangler, Kathleen, Electronic Meetings: Technical Alternatives and Social Choices (Menlo Park, California: Addison-Wesley Publishing Company), 1979. Karp, Peggy M., "Origin, Development, and Current Status of the ARPA Network." COMPCON 73, Digest of Papers, Seventh Annual IEEE Computer Society International Conference, 1973. Also reprinted in Blanc, Robert P. and Cotton, Ira W., (eds) Computer Networking (New York: IEEE Press), 1976. Lederberg, Joshua, "Digital Communication and the Conduct of Science: The New Literacy." Proceedings of the IEEE, 66:11 November 1978, 1314-1319. Lingwood, David A., Interpersonal Communication, Research Productivity, and Invisible Colleges. A Dissertation at Stanford University Department of Communication. August 1969. Lykos, Peter, (ed) Computer Networking and Chemistry (Washington, D.C., American Chemical Society), 1975. Martin, James, Future Developments in Telecommunications. (Second Edition) (Englewood Cliffs, New Jersey: Prentice-Hall, Inc.) 1977. ---------, The Wired Society. (Englewood Cliffs, New Jersey: Prentice-Hall, Inc.) 1978. Mathison, Stuart L. and Walker, Philip M., Computers and Telecommunications: Issues in Public Policy. (Englewood Cliffs, New Jersey: Prentice-Hall, Inc.) 1970. Mullins, N. C., "The Distribution of Social and Cultural properties in informal communication networks among Biological Scientists." American Sociological Review, 33 786-797. ----------, "An Organizational Approach to informal Communications among Scientists." Paper read at meeting of Eastern Sociological Association. Paisley, William, "Information and Work." Appears in Voight, Melvin and Hanneman, Gerhard, (eds) Progress in Communication Sciences, Volume II, (in press) March 1980. Price, Derek J. de Solla, Science Since Babylon (New Haven and London: Yale University Press), 1961. ----------, Little Science, Big Science (New York and London: Columbia University Press), 1963. Rice, Ronald E., "The Impact of Computer-Mediated Organizational and Interpersonal Communication." 221-249. Roberts, Lawrence G., "The Evolution of Packet Switching." Proceedings of the IEEE, 66:11 November 1978, 1307-1313. Spiegel-Rosing, Ina and Price, Derek de Solla, (eds) Science, Technology, and Society: a Cross-Disciplinary Approach (London and Beverly Hills: Sage Publications, Inc.), 1977. Turoff, Murray, "Human Communication via Data Networks." Computer Decisions, January 1973. Also reprinted in Blanc, Robert P. and Cotton, Ira W., (eds) Computer Networking (New York: IEEE Press), 1976. ---------- and Hiltz, Starr Roxanne, "Meeting through your Computer." IEEE Spectrum, May 1977, 58-64. Williams, Ederyn, "Teleconferencing: Social and Psychological Factors." Journal of Communication, Summer 1978, 125-130. Ziman, John, The Forces of Knowledge (Cambridge: Cambridge University Press) A P P E N D I X A The Finger Controversy A P P E N D I X B The Editor Discussion A P P E N D I X C A Map of the ARPANET