Path: utzoo!utgpu!news-server.csri.toronto.edu!mailrus!uwm.edu!zaphod.mps.ohio-state.edu!wuarchive!udel!princeton!phoenix!pucc!PSYCH@TCSVM From: harnad@phoenix.Princeton.EDU ("S. R. Harnad") Newsgroups: sci.psychology.digest Subject: PSYCOLOQUY V1 #8 (Electronic Archive : 288 lines) Message-ID: <9005312223.AA01110@reason.Princeton.EDU> Date: 26 May 90 00:37:15 GMT Sender: VMNNPOST@pucc.Princeton.EDU (Listserv to Netnews Gateway) Organization: Listserv to Netnews Gateway at pucc.Princeton.EDU Lines: 283 Approved: PSYCH@TCSVM PSYCOLOQUY Fri, 25 May 90 Volume 1 : Issue 8 Comments on Gardner's Electronic Archive by Stodolsky Response from Gardner ---------------------------------------------------------------------- From: david%harald.ruc.dk@pucc Subject: Archiving secure interactions This is the text of: Stodolsky, D. (in press). Archiving secure interactions [Letter]. Psychological Science. My invited comment on: Gardner, W. (in press). The electronic archive: Scientific publishing for the 90's Psychological Science. (Gardner, W. [wpg@virginia.edu], [Apr 1, 1990]. And could equally well be considered a response to the MegaJournal concept. Archiving secure interactions Gardner (1990) has done a fine job of illustrating the potential of computer mediated communication for scientific publication. However, there are both organizational and efficiency reasons for questioning the appropriateness of the model chosen. The model presented seems more appropriate for a student of the discipline than for a researcher. That is, there are (at least) two ways to use to use computer mediated conferencing. One is the archive or book approach, and the other is the meeting or argument approach (Stodolsky, 1990). With the archive or book approach, the user wants the received wisdom, presented in as clear and noncontroversial manner as possible. With the meeting or argument approach, the user wants to develop his/her reputation, and will therefore seek out errors and controversial material so that a response can be contributed (before it is done by others). It may be too strong a criticism to say the model is inappropriate, since it is actually the way of approaching the exact same information which is crucial. What will be argued here is that the great- est benefits will occur from the extension of computer mediation to include and in fact, restructure the process of peer review. Gardner (1990, abstract) argues that "the primary advantage of electronic publishing is not the inexpensive delivery of text, but the use of a central- ized archive to concentrate resources for discovery and utilizing information. The archive would . . . embodying knowledge . . . about . . . the intellectual goal of individual users." I will argue that the quantitative reduction in distribution costs is so great that it will lead to fundamental qualitative changes in scientific communication. And that centralization of resources is antithetical to a fundamental principle of science. Further, a system embodying intellectual goals of the individual user would best, for security reasons, be decentralized. Finally, a decentralized system would more easily make effective use of a scarce scientific resource (the judgments of reviewers). This is not to say that the proposed archive is a bad idea, in fact, it may be a pre- condition to the development of an improved system for peer review, and it would certainly be an element of such a system. The qualitative changes which can result from reduced distribution costs can perhaps best be illustrated by a concrete example. I will briefly describe my experience with the creation of a conference for the discussion of groupware on Usenet. Usenet is a world wide network of computers that participate in the distribution of network news, or just "news" to the initiates. There are about 10,000 computers on this network and about a million users. Some of these machines are large mainframes with hundreds or even thousands of users. Others are personal computers, often placed in the user's home, which serve a single person. Connection to this network requires merely freely available software and an agreement with the operator of a nearby machine to exchange news, typi- cally by using a modem and telephone. The news system functions as a distribut- ed bulletin board. Articles posted on any machine are distributed to all parts of Usenet within a few days, with most machines getting new articles in less than a day. The original function of this network was to exchange new software. This led to creation and distribution of "newsgroups" for discussion about the software. While computers and software remain a dominant topic, there currently exists about 500 different newsgroups covering a wide range of topics. One of the most widely read groups distributes jokes. This group is part of the "rec" hierarchy (groups oriented towards hobbies and recreational activities). Other Usenet hierarchies are "comp" (computer topics), "sci" (relating to the established sciences), "soc" (addressing social issues and socializing), "talk" (debate oriented), "news" (concerned with the news network and software), and "misc" (addressing themes not easily classified). One of the unique aspects of Usenet is that each machine or small group of machines is controlled by a separate authority. This autonomy is important in terms of the power relations most desirable for peer review. The creation of a newsgroup involves a discussion and voting process, the main objective of which is to determine the level of interest in a new topic and a proper name. A pair of newsgroups are available for these activities. Under the guidelines in force at the end of 1989, the process could be completed in about a month. Creation of the groupware newsgroup, comp.groupware, took about six months, due in part to my having to learn how to conduct votes. The total effort involved was roughly that of getting a journal article published. According to Reid (1990), comp.groupware had a readership of about 15,000 persons after its first month of operation. Average traffic was about three typewritten pages (six thousand characters) per day. The average cost per reader was estimated at two cents per month. (This figure includes only direct communication costs.) I draw a number of conclusions from this experience. First, the notion of a journal will have to be revised to fit the network environment. Second, the economic rational for the editor as gatekeeper in selection of material suitable for wide distribution has been eliminated. Self publication will become common and selection of material will become the responsibility of reader. This, of course, assumes that centralized quality control is best eliminated. An ideal within science is that only the logical force of an argument should control its influence. Power relations inherent in the operation of a peer review journal contradict this ideal (Michel, 1982) and sometime lead to irresponsible behavior (Garfield, 1988). This contradiction can avoided if a decentralized system for quality control is created. Computer networks open this possibility, and in fact, make available superior methods for quality control. Stodolsky (1990) points out that a loss of information occurs in peer review because of the use anonymity to protect referees and to ensure impartial review of manuscripts. This loss can be avoided if pseudonymity is used. A pseudonym system would permit scientific reputation to be developed as easily by reviewing other's articles as by publishing ones own. A system which makes effective use of peer review judgments both in the selection of manuscripts and in the development of reputation is most practical and secure with decentralized computer capabilities. A secure pseudonym system requires a powerful and sophisticated technology. However, if the goal of a system embodying knowledge about the intellectual goals of individual users is to be realized, it is unlikely that can be achieved with anything less. In today's highly competitive environment, authors often will not risk exposing their intellectual goals prematurely. Even if this data is retained on the user's computer, interactions with an archive could expose them. Also, since a close coupling between archived data and the intellectual goals are desired, the separation of these types of information could lead to inefficiencies. Recent research suggests that an integration of centralized and decentralized storage mechanisms leads to the most desirable overall capabilities (Gifford, 1990). Thus, both security and efficiency concerns, as well as the dynamics of social organization suggest that maximal decentralization is appropriate. Gardner (1990) proposes formation of an organization run by scientists to support users of an electronic archive. A redirection of organizational resources already in place may be more appropriate. Resources that would otherwise be consumed in creating a new organization can better be used to promote the integration of existing organizational units and individual scientists into the network environment. For a start, every society office with responsibility for publication should be provided with resources and training to operate a node on the Internet. For each publication released, at least an abstract of the articles contained in that publication should be directed to the appropriate Usenet newsgroup. The Bitnet Psychology Newsletter and Bulletin Board has been restructured and is now available both by electronic mail and as the Usenet newsgroup, sci.psychology.digest. This could be an appropriate channel for the publication of this type of material. The support provided to this publication should certainly be continued. At least one node should have responsibility of archiving newsgroups of interest to psychologists. Another could make material available to those without access to the network or interest in using it directly. The best service can be probably be provided by a decentralized system, where staffs of one or two persons work under the direction of individual scientists. Savings from this strategy could be invested in applied research directed towards improving the quality of software available for news reading and for supporting peer review. Usenet and many other networks are currently dominated numerically by computer professionals and students whose social science knowledge is vanishingly small. While a powerful mechanism for the transport of news is in place and working remarkable well, considering the social environ- ment of the network, the best news reading software sometimes does not even present target articles before responses to them. There exits a fertile field for the application of psychological knowledge, and real opportunities for social scientists. For example, support for finding the places in discussion where a person could have maximal impact are needed. Psychological theories could be embodied in news reading software in order to facilitate this type of search. Such activities would provide benefits to users and to researchers. This approach would also prevent the isolation from mainstream developments in electronic publishing and the wider community of on-line users. In summary, current organizational resources could be restructured to support on-line users. This could probably be undertaken with little additional fund- ing. The next step would be to add editors and then editorial boards to the network. At the same time, a number of independent researchers could be developing software for making utilization of the expanding network easier and more effective. The outputs of this research could be made freely available to society members. The result would be a rapidly expanding network of users forming an closely integrated and quality conscious scientific community. The electronic archive would play a key role in this organizational restructuring, but without such restructuring the contribution made by such an archive might not be appreciated. References Garfield, E. (1988). Refereeing and peer review: Part 1. In Garfield, E. Essays of an information scientist: Towards scientography (Vol. 9). Philadelphia: ISI Press. Gifford, D. K. (1990). Polychannel Systems for Mass Digital Communication. Communications of the ACM, 33, 141-151. Michel, F. C. (1982). Solving the problem of refereeing. Physics Today, 35(12), 9;82. Reid, B. (1990, February, 1). USENET Readership report for Jan 90. Usenet conference news.lists. (Message-ID: 382@apostrophe.dec.com) Stodolsky, D. (1990). Protecting expression in teleconferencing: Pseudonym- based peer review journals. Canadian Journal of Educational Communication,19(1). ([1989, May 9]. Communication Research and Theory Network [CRTNET],No.175 [Semi-final draft available by electronic mail from LISTSERV@PSUVM.BITNET at University Park, PA: The Pennsylvania State University, Department of Speech Communication and COMSERVE@Vm.ecs.rpi.edu at Troy, NY: Rensselaer Polytechnic Institute, Department of Language, Literature, and Communication]) David S. Stodolsky Internet: david@ruc.dk Department of Psychology IP No.: 129.142.144.20 Copenhagen Univ., Njalsg. 88 Voice: + 45 31 58 48 86 DK-2300 Copenhagen S, Denmark Fax: + 45 31 54 32 11 ------------------------------ From: "William P. Gardner" Subject: Response to Stodolsky by Gardner Here are some reactions to David's perceptive comments about my article "The electronic archive: Scientific publishing in the 90's" (in press, _Psychological_Science_ -- anyone interested in a copy should write to me at wpg@virginia.edu, please specify ASCII, LaTeX, or surface mail.) * David contrasts the archive approach with a meeting or argument approach. As suggested in my article, the two approachs are complementary rather than competing (see the vignette about user B). Moreover, I believe that network discussions will be most fruitful if there are electronically archived references to serve as fixed of points of reference for the discussion. * I am sympathetic to many of your comments about the problems of centralization (I have received a lot of critical mail on this point). The idea in the archive (a kind of Bibliotheque Nationale out on the net) may take centralization farther than it needs to go. Still, I believe that the archive or library concept is important. There needs to be some single, coherent scheme for organizing the material so that each electronic article exists as a permanent, efficiently retrievable record. One way to do this is to store all electronic texts in the machines of one institution (a Bibliotheque Internationale? :-) ). But this could place too much power in one institution. A better idea may be to design a network-wide hypertextual reference scheme, which would give a netaddress and filing number for any article on any computer (the way Library of Congress numbers work for libraries.) Then anyone can start a journal by putting an archive on-line. Nevertheless, somebody somewhere needs to be copying it all onto optical disks and putting it into vaults, so that the articles survive when Journal X goes out of business. A centralized cataloguing service may also be required. So again, we need some central organization. * I am not enthusiastic about the USENET model. It was, in fact, disatisfaction with News that led to the archive paper. Even more importantly, I don't believe that minor restructuring of current resources will do the job -- the job is much too big. Current electronic journal models (including USENET) will not work because the flat ascii file won't work. We need hypertext and professional standard typography, and anything else is a step back from the utility available from print technology. * Your ideas about a new peer review system are fascinating and deserve the widest possible discussion. In summary, I am inclined to believe that the ideal scheme falls somewhere between the position staked out in the target article and what you suggest in this letter. Your ideas about using the encryption technology to establish secure net identities may be an essential component of the solution. ------------------------------ PSYCOLOQUY is sponsored by the Science Directorate of the American Psychological Association (202) 955-7653 Co-Editors: (scientific discussion) (professional/clinical discussion) Stevan Harnad Perry London Psychology Department Dean, Graduate School of Princeton University Applied and Professional Psychology Rutgers University Assistant Editors: Malcolm Bauer John Pizutelli Psychology Department Psychology Department Princeton University Rutgers University End of PSYCOLOQUY Digest ******************************