bannon@sdcsla.UUCP (Liam Bannon) (10/14/84)
I have not finished my study of software support for person-to-person interaction, but decided that I should send out a note on the results of my query on "computer-mediated social interaction" to let you know what I have learned to date. First of all, a public thanks to all those who responded to my query, I have already thanked most respondents privately. To conserve space (and requests for anonymity in some cases!), I do not identify the sources of the quotes below. -liam bannon ------------------------------------------------------ Topics: General Use of Computers - Asynchronous Communication Facilities Synchronous Communication Facilities - Miscellaneous General Use of Computers: Some people noted how just having a computer around seemed to stimulate cooperation, "Although we have only a few general-use data bases in the customary sense, there is a considerable amount of data sharing in many forms - made practicably possible only by our super-mini-based computer system. And, programming packages for various applications (e.g. signal processing) often get developed for and by one group, then are applied by many of the others. Proposals and papers, more commonly than not, are prepared jointly with multiple versions of drafts being revised and edited back and forth between the various researchers in different fields. ... people from a variety of different disciplines that, only a few years ago wouldn't have though about sharing data, findings and working cooperatively together are, in fact, doing that now. All despite the fact that we have a less-than-rudimentary electronic mail system on our main system, and none of the nifty tools that supposedly provide for handy electronic interaction." Asynchronous Communication Facilities Here, I include such facilities as electronic mail, what is now called computer conferencing, and forms management. Electronic Mail: Most computer systems now have some form of electronic mail, and their benefits have been discussed often in the literature, so I don't propose to say much about them. Most of the systems also allow for reasonably large files to be passed around, thus allowing for several people to work on a paper. If there can be some sort of shared file system, so much the better, e.g. "At Xerox, we used electronic mail and shared electronic filing to great advantage -- long-distance cooperation on documents was the rule, since about 2/3 of our people were in El Segundo and most of the rest in Palo Alto. I wrote a couple of papers this way, with co-authours down south; Star's Functional Description was maintained similarly;... the general approach was to keep the text in a public spot; give one author at a time the write access (often, divvy it up and let different folks be working on different parts; the "write access" was by consensus, rather than by adjusting privileges -- the privilege structure certainly would have supported it, but for papers, we didn't bother. For code, it was another matter -- that got formally checked out and back in again.) comment & revise via electronic mail, with occasional check-ins where people got a new consistent version." Another person commented on how important the ARPAnet was in the development of the Ada language, "Not only were the design and review processes mediated by the ARPAnet, but the language design team was geographically distributed. Jean Ichbiah and several others in France, a group of key people in England, the administrative work and design of a test compiler in Minnesota, and several other key people in the United States, Germany, the Netherlands, etc." [My own use of netnews to gather information on this topic is another example of how valuable communication networks can be - Editor] Computer Conferencing: This term now refers to computer systems that provide additional facilities for keeping a record of all messages on a topic, allowing one to set-up conferences with certain people, and browse through the topics and messages on each topic. There are usually facilities to send public and private messages, and to find out information on the conference participants. The use of the term "conference" is a bit misleading, as usually we refer to conferences as being in real-time, namely all the participants being active at the same time, whereas this is NOT the case in computer conferencing. There are a number of well-known conferencing systems available, including the New Jersey Institute of Technology's EIES, the Univ. of Michigan's CONFER, the PLANET/FORUM system (Institute for the Future), and the Swedish COM and PortaCOM system. Most of these systems have been studied intensively by means of on- and off-line questionnaires given to the participants, analyses of usage statistics on the system, and particpant observation. In reply to my query, there was a very long bibliography of reports produced by the EIES group that was broadcast over the net, so I will not include it here. To get a feel for the kind of studies that were done, see Starr Roxanne Hiltz's book "Online Communities" ( New Jersey: ABLEX, 1984). Synchronous Communication Facilities This is what I was really interested in. Here we are talking about people communicating in real-time through a computer system. Many systems have some simple means of directly "writing" or "sending" a note to another terminal...often used within a building for short announcements like "want to go to dinner now?"!!! Generally people do not carry on any sort of sustained dialogue this way as often the line is only half-duplex, so you have to wait until the other person has finished typing before you can respond. Some systems now split the screen and allow simultaneous typing, which makes an order of magnitude difference in the perceived utility of the system, but this is all that is allowed, ie. you can't get access to another person's files, or see their screens, etc. A much more powerful tool is a facility like the link command on the TENEX system, which allows one to "link" to another tty and actually see what is on the other's screen, and have access to their programs, etc. Here are some comments on the usefulness of such facilities. (a long, but striking example) "We have a junior programmer working on the project I am "leading" [the programmer is located 2,000 miles away from the project leader who is reporting this incident! - Editor]. We are working on a Tenex-like system. I wanted to find out how things were progressing and I knew that there was a problem with bad hardware that made the terminal line in her office unusable, so I figured she was probably in someone else's office. So instead of calling on the phone I found her on the system and linked to her. It came out that there was a problem with a piece of code she was working on. It was giving an error message that a certain record wasn't in the database, even though it was. She couldn't find anyone in the office who could tell her how to figure out what was wrong (it wasn't the obvious things). So I converted the link into an ADVISE link. (... tenex ADVISE puts the "advisor" into a state where his keyboard input is put into the input stream of the advisee, as well as the shared output function of LINK.) Then I told her (via comment command) to show me the right source things. She jumped to the right code on the screen and I then gave some more commands to jump around to other procedures being called, etc. Didn't find anything obviously wrong. So I told her to get the thing running under our debugger and put a breakpoint at certain place and do whatever it was you do to get that code to be executed. So she loaded the program and activated the debugger and went into the program and gave a command which resulted in the breakpoint being hit. Then I started giving debugging commands and scouted around in the executable image and eventually displayed the filename associated with a certain "statement identifier" (never mind what that is..). As soon as I did that she said (typed) "Oh, I know, ..." etc. And also that she didn't understand some of the debugger commands I gave. So I told her what I had done and why. And now she knows what needs to be fixed... and we broke off with gossip, goodbyes, etc." "The ability to link terminals seems to be of most use for remote demos, instruction, or receiving expert assistance. I know of lots of people who are strongly in favor of it for these uses and I have seen it used effectively in these contexts." "When I was user support for the Datacomputer in Cambridge more years ago than I like to think, some of our biggest users were out in California, coming in over the ARPAnet. As expected, there were occasional problems, which mail & long distance phone calls didn't really ease. Better was for one of us to log in on the other's machine and run the offending code with the other one linked on, occasionally suggesting debugging probes to try or alternative strategies." "TENEX's and TWENEX's LINK - By far the best. This system links two terminals together; not only can one hold conversations, one can also display files, demonstrate how to do this or that, give demonstrations, etc. VMS's PHONE - More features, less useful. This sets up two windows (more, if needed) on your screen, and your text is scrolled by in these win- dows. You can also display files in it. It's better in that it keeps your text seperate from the other - so that if you both type something at once, it isn't garbled - but you can't do anything else (like run programs, etc.). My Ideal - I'd say T[W]ENEX's LINK is the best approach, though it would be nice to have a windowing PHONE-like tool to go with it for the chatter." Perhaps the most persistent researcher in the field is Doug Engelbart, who has been concerned about providing support tools to knowledge workers for more than 20 years, initially at SRI, where he developed the NLS system, and subsequently at TYMSHARE with the re-named AUGMENT system. This system has explicit support for synchronous interaction. The AUGMENT Conference system "permits a user to call an on-line conference of two or more people, view and edit files, add and remove conferees, pass the gavel, and transparently connect to other machines via TYMNET or ARPANET. Televiewing is usually done in conjunction with a telephone connection, and is often used to support document review and revision in a synchronous mode, where all conferees can see and discuss changes as they are made" (Engelbart,1984) [For an overview of the Augment system, see the Seybold Report on Word Processing, October 1978. For an overview of some of the collaborative support tools in AUGMENT, see the paper by Engelbart "Collaboration Support Provisions in AUGMENT", in 1984 AFIPS Office Automation Conference Proceedings.] There has recently been a renewed interest in the area of support tools for collaboration. A recent workshop on the topic of "Computer-Supported Cooperative Work" was organized by Irene Greif (MIT) and Paul Cashman (DEC), and a report on this workshop should be available shortly. At the Laboratory of Computer Science at MIT, GReif and others have a project in the area of synchronous communication facilities (See "Software for Interactive Online Conferences", S. Sunar and I. Greif, 1984 ACM - SIGOA Conf. on Office Information Systems). Several other groups around the country are starting projects in the area, at Univ. of Maryland, Xerox PARC, DEC. Miscellaneous There is a lot of work on tools for document creation, retrieval and manipulation that I believe is relevant to the provision of software support for collaboration which could be mentioned. Here are a couple of references that I feel are of interest: R.H. Trigg &M. Weiser. TEXTNET: A network-based approach to text handling (Dept.of Computer Science, Univ. of Maryland, 1983(4?)) D.C. Engelbart. Authorship provisions in AUGMENT (Proc. 1984 COMPCON Conf.) D. Lowe The representation of debate as a basis for information storage and retrieval (AFIPS NCC Conf. July,1984) Xerox PARC - the NOTECARD and Annoland projects. Sorry for the length of this message. If you have any comments, additions, etc., please let me know, as I am writing up a paper on this area, and would like to mention any other relevant projects and facilities. -liam bannon.