|Author:||Jeffrey R. Yost|
|Title:||Overcoming the Discipline Divide: Knowledge and the Advancement of the History of Software|
|Publication Info:||Ann Arbor, MI: MPublishing, University of Michigan Library
This work is protected by copyright and may be linked to without seeking permission. Permission must be received for subsequent distribution in print or electronically. Please contact firstname.lastname@example.org for more information.
Overcoming the Discipline Divide: Knowledge and the Advancement of the History of Software
Jeffrey R. Yost
vol. 4, no. 1, April 2001
Overcoming the Discipline Divide: Knowledge and the Advancement of the History of Software
This study briefly relates both the historic and contemporary disparity between the use of networking technology to advance scientific and engineering fields relative to areas in the humanities and the general failure to capitalize on interdisciplinary research opportunities made possible by the Internet. It refers to these developments as the discipline divide. The body of article presents a case study of a NSF-sponsored (NSF-9979981) knowledge networking project underway at the Charles Babbage Institute (CBI) to research and produce Web-based infrastructure resources to further the historical study of software. It indicates how this project, by combining a committee structure and network communication between historians and technical experts from the software community, provides a potentially valuable model for research in the recent history of science and technology, as well as one possible method to help mitigate the existing discipline divide.
As the use of networking technology becomes increasingly ubiquitous in commerce, finance, education, research, and many other realms, concomitant concerns about access to and underutilization of networking resources abound. Internet access is fundamentally an ethical issue that demands new policies and initiatives to begin to narrow an ever-widening "digital divide."  While the divide in access to and use of information along racial and economic lines is clearly the most critical and immediate concern, the Internet creates or reinforces many different types of divides. One such divide, which I refer to as the "discipline divide," exists between scholarly fields. This takes two primary forms. First, it perpetuates communications by scholars, students and other individuals within traditional discipline boundaries. Second, historical and other factors facilitate more extensive and involved use of networking resources in some fields than in others. The first is not a product of networking, but rather, a failure to fully take advantage of the Internet to engage individuals and create and distribute information across disciplines. The latter, extends from the origins of the ARPANET, CSNET, and NSFNET as a government-funded projects to advance science by the Department of Defense Advance Research Projects Agency (ARPA), the National Science Foundation (NSF), the National Institutes of Health (NIH), and other federal agencies.
From SUMEX-AIM, an NIH-funded networking project begun at Stanford University in the early 1970s to connect leading research centers in the United States studying artificial intelligence applications in medicine, to the continually advancing on-line inter-firm computer-aided-design (CAD) and computer-aided-manufacturing (CAM) at many industrial companies of the past decade, computer networking has become an extremely important scientific and engineering tool. 
Scholars in fields outside of engineering and the sciences have made a number of attempts to integrate and expand the role of computer networking in their disciplines. These efforts have not been a central focus within the humanities, despite dedicated work in this area by a small number of individuals and organizations both nationally and internationally—including, the American Association for History and Computing (AAHC). Overall, accomplishments in this area have been significant, but uneven. The greatest achievements have been in the distribution of existing information and education efforts,  rather than the creation of new documentation and research. This has clearly been true in the field of history.
Andrew McMichael, in an article in the American Historical Association's (AHA) Perspectives from early 1998, identified some of the key problems in current applications of networking technology to the field of history. He stated that Web sites often contain inaccurate or undocumented information, that Usenet newsgroups are frequently dominated by "buffs" where those "who repeat their arguments the most and the loudest tend to dominate the discussion," and H-Net, the Humanities and Social Sciences Network of more than 90 Listservs, tends to be "little more than a means to avoid using the Online Computer Library Center (OCLC)" and other reference sources.  Notwithstanding these largely valid critiques, McMichael tends to downplay the fact that there is some significant information on the Web, that if viewed with a critical eye, can be helpful to educators and students of history. Resources such as Dennis A. Trinkle, et al.,'s The History Highway: A Guide to Internet Resources demonstrate the current breadth of history Web pages/Internet sites, and provide guidance on what various sites offer. 
McMichael also neglects a fundamental deficiency in the current state of computing and history: the lack of communication between scholars in different disciplines and the lost opportunity for those studying recent history to enter into an electronic dialogue with some of the actors of the events and institutions they are studying. An area in which this could be particularly valuable is the recent history of science and technology. Such a dialogue could potentially lead to the development of important new documentary resources in targeted sub-disciplines of the history of science and technology, as well as the identification and possible collection and preservation of materials at appropriate archives repositories.
Ironically, as the digital world of the new century advances possibilities for storing, communicating, and processing information, it simultaneously creates new challenges in preserving and providing access to data that originates and is stored in a variety of formats. In the history of science and technology, laboratory notebooks, minutes of meetings, project reports, correspondence and other primary source materials are increasingly produced in digital form and often are not retained as long as their hard copy counterparts of the past, or they become inaccessible through obsolescence resulting from changing hardware, software, or systems. Thus, they generally are less likely to make their way to repositories that collect and provide access to such records. The accelerating amount and importance of science and engineering work within corporate R&D laboratories, and at universities partnering with corporations, has led to escalating concerns over protecting proprietary knowledge and intellectual property.  Information on the Web is used more frequently than paper records of archives repositories and is much more easily copied, altered, redistributed, etc.  Proprietary issues aside, archives will likely be faced with escalating technical and economic challenges of long-term preservation of records in a variety of formats.
In addition to educating individuals and organizations that generate significant records documenting the sciences and engineering, and developing the means and standardized methods of long-term preservation of records at archives repositories, there is a need for greater collaboration between historians, and scientists and engineers to take a proactive role in the creation of documentary materials. In short, there is need for using networking resources to reduce the discipline divide between scientific and engineering fields and the specialists in the history and philosophy of science and technology attempting to expand humanistic understanding of scientific/technical areas of research and development.
.02. BUILDING A FUTURE FOR SOFTWARE HISTORY
I will now turn to a short case study of the Charles Babbage Institute's (CBI) NSF-sponsored $488,000 three-year project, entitled, "Building a Future for Software History" (NSF-9979981). For those unfamiliar with CBI, the Charles Babbage Institute is an international research center and archives for the history of information technology. CBI's archives collection contains more than 6,000 cubic feet of primary source material in the history of information technology, including: corporate records, papers of leading computer scientists, trade organization records, product literature and manuals, oral histories, and many other types of documentation. The Institute conducts original research in the history of computing, software, and networking and provides fellowships to doctoral students working in these areas. I serve as the Principal Investigator of CBI's NSF-sponsored software history project.
"Building a Future for Software History" is an experiment in knowledge networking that brings together technical experts and historians to development documentary resources in the history of software. While there have been steady advances in the collection of materials and analysis of the history of computing hardware, aided in large part by work at or facilitated by the Charles Babbage Institute, the history of software has been ignored, both in terms of historical examination and the systematic preservation of documentary records. In 1997, leading industrialists and academicians in computing and software who serve as Trustees of the Charles Babbage Foundation (CBF), our Institute's funding arm, identified and encouraged CBI to develop a project to advance the history of software. Benefiting from a report of the CBF Software History Task Force led by George Glaser,  Robert W. Seidel and I co-wrote a proposal, which consists of four primary initiatives: 1) creating a knowledge network of individuals from the software community, 2) using this network to produce an authoritative on-line software dictionary, 3) an oral history initiative to interview pioneering software developers, and 4) an electronic journal of software history. In addition to myself and Professor Seidel, Project Manager and CBI Post-Doctoral Fellow Philip L. Frana, CBI Director Arthur L. Norberg, and CBI Archivist Elisabeth Kaplan are serving on this project. 
I will now expand upon each of these initiatives, but focus primarily on the knowledge network and creation of the on-line software dictionary. I will address early lessons we have learned as we enter the second year of the three-year project, and place developments within the context of other attempts to use computer networking technology to advance the documentation and understanding of the history of science and technology.
Current literature in the history of software is sparse and lacks critical understanding. While there is a new book on Bill Gates or Microsoft on almost a weekly basis, these works seldom analyze fundamental technological and managerial issues involved with software development. Scholarship addressing software research at academic institutions and at personal computer software companies is unusual, while studies of software for mainframe and mini-computers is even less common. Thus, in the first three-year phase of our project, CBI chose to focus only on software in the mainframe and mini-computing eras. This will also provide important background for future work to understand the development of software in the PC era and beyond. 
During the mainframe and mini-computing eras, software programming was particularly craft oriented, and much software was programmed by users, academicians, and relatively small software service firms. Despite the geographical and organizational dispersion of activity, software development remains a cumulative enterprise. Portions of code are reused and early decisions create path dependencies that simultaneously open and constrain future possibilities. User groups, such as SHARE and DECUS, for IBM and Digital Equipment Corporation computer users respectively, were fundamental to this process through the creation of libraries of code that were shared among members.
As a first step in making the history of software possible, and at the advice of the CBF Software Task Force, we concluded that it would be necessary to gain a better grasp of the terminology of the discipline; in other words, develop a clearer understanding of the changing nature, meanings, and institutional contexts of software technologies and techniques over time. Extensive analysis of the existing historiography in the history of software, technical articles, as well as scholarship presented at software history conferences co-sponsored or sponsored by the Charles Babbage Institute over the past year in Paderborn, Germany and at Xerox PARC in Palo Alto led to and reinforced our conclusion that identifying the terminology and understanding its contexts would be critical to communication and furthering this area of study.
To advance understanding of software history terminology, the first two initiatives of our project are to build a knowledge network of distinguished figures contributing to early software development, and to engage these individuals in an electronic dialogue to produce an on-line dictionary of software history. This dictionary is designed to be accessible to educated non-specialists and will serve as an authoritative resource for computer scientists, industrialists, historians, students, and others.
.03. THE KNOWLEDGE NEWTWORK AND SOFTWARE DICTIONARY PROJECT
Last year, after conducting extensive research in trade and scholarly publications and communicating with a number of individuals in computer science and industry, CBI formally designated nine software categories for the dictionary project and identified the most qualified individuals to invite to serve as committee members and chairs. The categories/committees include: operating systems, database software, graphics, programming languages, programming techniques, software engineering, scientific applications, business applications, and networking. The next step involved the formation and operation of a pilot committee to identify best practices.
In forming the pilot committee we sought the most distinguished individuals in the field of operating systems from industry and academe. Approximately 40% of individuals asked to participate agreed to serve on the pilot Operating Systems Committee. As expected, the percentage of individuals signing on from the academic community was higher than from industry, but enthusiasm for the project and willingness to participant was present in both sectors, and the distinction between them increasingly is blurred by the consulting practices of many top computer scientists.
The highly accomplished individuals on the Operating Systems Committee include: Bernard A. Galler, University of Michigan (Chair); Ashok K. Agrawala, University of Maryland, College Park; Michael J. Alexander, Arbortext; Marc A. Auslander, International Business Machines; Edward G. Coffman, Jr., Columbia University; Peter J. Denning, George Mason University; Alan Cary Shaw, University of Washington; Avi Silberschatz, Bell Laboratories; Andrew S. Tanenbaum, Vrije Universiteit, Netherlands; Thomas H. Van Vleck, Encirq, Inc.
CBI has recently begun the launch of the Database, Graphics, Scientific Applications, Programming Techniques and Programming Language Committees and project staff members are pleased with the acceptance rate of about 60%. We attribute this increase to the growing momentum of the project and the fact that individuals invited to participate on these other committees have been almost universally impressed with the stature of those serving on the Operating Systems Committee.
The Operating Systems Committee has developed the criteria for inclusion, and designated more than 120 of the most important terms in operating systems. They currently are in the process of choosing the terms in their particular area of expertise and writing draft entries, either individually or collaboratively. As these terms were introduced by committee members, an extensive online debate took place about what components were critical to constituting an operating system. This discussion ranged from the broad to the specific, defining operating systems as managers of resources that run on the "raw hardware," but exploring some of the critical distinguishing features as well, to set them apart from languages and other types of software. Important distinctions were made to exclude interpretive programming languages running in user space such as FORTH and JAVA, despite the fact that these and other languages provide a "virtual machine," one of the key features of an operating system. While excluding characteristics were identified, essential criteria for operating systems were also put forth, such as the ability of operating systems to have all or part of the software running in a "privileged mode." We are extremely pleased with the initial results of the pilot Operating Systems Committee, and the process appears to be working as planned.
More than two dozen entries have been completed. Once 200 are completed and validated by the various committees, CBI will launch the on-line searchable Software Dictionary. The defined terms will be hyperlinked whenever they appear in the text of other entries. CBI Archivist Elisabeth Kaplan is in the process of designing structure and features of this site that we expect to launch by Fall 2001.
.04. ORAL HISTORY INITIATIVE
We have also begun conducting thirty-two oral histories with software pioneers, which the Institute records, transcribes and places on the CBI Web site. The Charles Babbage Institute has conducted more than 200, and holds more than 300, oral histories with leading figures in the history of information processing. While oral histories have a number of shortcomings and must be used with a critical eye, the extensive preparation and careful selection of subjects, have resulted in research-grade oral histories that serve to document events and circumstances in the history of information processing that would be difficult or impossible to research in any other way.
The current initiative extends this mission, with the targeted focus on the history of software. The oral histories work synergistically with the development of the online software history dictionary, with both informing each other along the way. The two initiatives will also help identify collection development opportunities in the software field for the Institute's archives.
Thus far, as part of "Building a Future for Software History," we have interviewed such distinguished individuals as Edward Feigenbuam, commonly credited as the father of Expert Systems; Douglas Hofstadter, the author of Godel, Escher, Bach: An Eternal Golden Braid and a leading figure in software and cognitive science among other fields, and Douglas Ross, the creator of structured analysis and a fundamental early contributor to applications of software to industrial automation. 
.05. ELECTRONIC SOFTWARE HISTORY JOURNAL
The electronic journal is nearing its launch, as we currently are determining its look, structure, policies, editorial board, etc. It will be a refereed journal (evaluation by at least two outside experts) that will offer continuous publication with volumes separated by calendar year. The articles will be published in both HTML and PDF formats. The journal will be interdisciplinary in nature and serve as an outlet for a broad range of topics and methodological approaches, including sociological and anthropological analyses of software development, the history of the software industry, historical studies of management within software firms, user programming, first hand accounts of developments, social impacts and ethical implications of software technology, etc.
Understanding the ephemeral nature of digital-only journals, hardcopies will be kept, and made available as part of the Institute's permanent archival collection. Additionally, all electronic correspondence is being printed on acid-free paper and will be stored in the archives and available to researchers at the conclusion of the three-year first phase of the NSF project. This will serve as a resource for researchers that may want to understand the nuances of the debates of committee members that resulted in the final online dictionary. Additionally, it will provide the grist for understanding the successes and failures of the methodologies employed by CBI in this experiment in knowledge networking to advance the infrastructure and understanding of a sub-discipline in the history of science and technology.
.06. IN CONTEXT: INFRASTRUCTURE NETWORKING PROJECTS IN THE HISTORY OF SCIENCE AND TECHNOLOGY
"Building a Future in Software History" is not unique in using networking technology to develop and disseminate research in the history of science and technology. Nevertheless, there are some significant methodological differences between this project and ongoing and past projects elsewhere that we believe will be critical to its success.
The CBI project differs from the work of most Special Interest/Use Groups in that there is a committed focus to interdisciplinary participation and a clear definition of deliverables. Use Groups, such as the 90 or more specialty groups on H-NET, rarely establish defined goals and infrastructure products, instead they exist as a forum to voice opinions, get initial feedback on potential research, locate a difficult to find resource, etc. At these tasks, the Use Groups appear to be fairly successful. H-NET, however, does not often benefit from participation from individuals outside of history, including scientists and engineers, groups that could add significantly to discussions in sub-disciplines such as the history of science and technology.
The CBI Software History Project is also not unique in establishing a set of topics and soliciting the on-line participation of scientific and technological experts to develop resources for researches to be placed on the World Wide Web. Different methodologies, however are being employed by the Institute, with the hope of yielding a more extensive and higher quality product.
The Sloan Foundation has funded an important project to build and disseminate information on a small number of designated technological projects, including the construction of the Boston Central Artery Tunnel, the history of programming languages and the development of the computer mouse.  These projects solicit participation over the Internet and without filtering and place the materials and viewpoints that people provide up on the Web. This strategy of inclusion has a number of benefits as well as some significant drawbacks. On the positive side, previously unidentified individuals that have important information or documentation may surface and offer important perspectives or resource materials. Additionally, historians or archivists involved in running these projects do not impose biases they may have on the future value of certain information, ideas, or records.
On the other hand, the level and quality of participation is left to chance. The result could be low participation or involvement from individuals on the periphery of events that have inadequate recollection or basis for documenting historical developments. Low participation or inaccurate information may lead potentially valuable contributors to decide against getting involved. Furthermore, there is no guarantee that contributors who are knowledgeable about the projects will give accurate representation of past events. The current state of the Web sites of these projects indicates that while some valuable knowledge has been gained, participation has been relatively low and the quality of information is mixed. This leads to the risk of the use, misuse, and proliferation of information that has not gone through any formal review process.
The Charles Babbage Institute's software history project seeks to either circumvent or limit these drawbacks, while making every effort to cast a wide net through soliciting the participation of a wide-range of leading figures in a particular area to serve as committee members. The committee members are then encouraged to utilize as many contacts as possible as resource people for providing information for the dictionary entries, and collections of documentation that might be appropriate to add to the CBI archives. This process extends the network further and gains momentum from the information and knowledge it produces.
The problem of inaccurate information up on the World Wide Web is widespread and poses significant challenges to researchers. By utilizing a committee of experts and all available resources at the Institute we are striving to produce an authoritative resource for computer scientists, the software industry, historians and scientists studying the history of software, and other interested individuals. The project takes advantage of the flexibility and speed that digital communication (group email accounts for each committee and a designated portion of our Web site) to provide information of ongoing committee work and resources developed by project participants. For instance, CBI project staff has produced a bibliography of more than 1,500 citations of significant articles and books with information on software development in the mainframe and mini-computing eras. This has been placed on the Web, to assist committee members and resource people in the production of entries. 
The dialog that CBI is helping to coordinate between distinguished individuals from industry and academe with unique knowledge of institutions, events, practices, technologies, and people in specialized areas of software history, along with the perspective, research, and oversight that the four historians of science and technology and the archivist on the project provide, represent a potentially important methodology for building an infrastructure in understudied areas of the recent history of science and technology. The ongoing research indicates the benefits of combining a traditional committee structure with networking technology to ensure quality participation.
More broadly, CBI's "Building a Future for Software History," provides one model for addressing the two fundamental aspects of the discipline divide: the disparity of technical and non-technical fields in incorporating networking technologies to advance research, and the lack of communication and interaction between scholars in different disciplines. Fields in the humanities that by their very nature are interdisciplinary, such as the history of technology and science, are particularly apt to benefit from greater use of networking tools to facilitate an ongoing discourse and the sharing resources between individuals and groups from technical communities engaged in historically significant projects and scholars who study the technology and context of such developments.
1. Kathy Koch. "The Digital Divide: Should Internet Access for the Poor be Subsidized? " Congressional Quarterly 10:3 (2000) 41-64. Alan Murray. "Who Wins in the New Economy? Web Empowers the Little Guy, Populists Say, But Critics See a Growing Social Gap," Wall Street Journal (June 27, 2000) B1. The term "access," as I use it in this paper, refers not only to access to an Internet connection, but also the readiness of access (home connection versus limited use at public library or school) as well as the basic knowledge needed to make effective use of the resource (e.g. point and click operations and tools, conducting and narrowing searches on search engines, the best search engines for particular types of searches, etc.). For many individuals, the latter factors prove to be greater obstacles to sharing in the educational and other opportunities offered by the Internet.
2. The Seeds of Artificial Intelligence: SUMEX-AIM (Bethesda, Maryland: Division of Research Resources, National Institutes of Health, 1979). While some cooperative networking projects in the humanities began in the 1980s (Guttenberg and Text Encoding Initiative), computer networking was and remains far less prevalent in the humanities than in scientific and engineering fields.
3. A survey by the American Association for History and Computing indicated the widespread use of the World Wide Web in teaching history by the late 1990s, indicated that research assignments were less common, and reported a substantial concern of educators regarding the accuracy of information on the Internet. Dennis A. Trinkle. "History and the Computer Revolutions: A Survey of Current Practices," Journal of the Association for History and Computing 2:1 (April 1999). http://hdl.handle.net/2027/spo.3310410.0002.107. Many historians have written of the potential benefits of Internet, but stress the importance of critical use of resources. For instance, see Michael O'Malley and Roy Rosenzweig. "Brave New World or Blind Alley? American History on the World Wide Web, " Journal of American History 84:1 (1997) 132-155; Sally Gregory Kohlstedt, et al. "Enhancing History, " Organization of American Historians Newsletter 26:2 (1999) 1, 7-9; and Corley, Julie A. "Can the Web Really Do It All? Perceptions of Historical Research on the Internet, " Public Historian 20:1 (1998) 49-57. On the other hand, some historians, for a variety of reasons, oppose the use of the Internet in higher education. David F. Noble, for instance, has argued that the movement to require use of on-line technology in higher education represents the culmination of the commercialization of academe since the 1970s. David F. Noble. "Digital Diploma Mills, The Automation of Higher Education, " Monthly Review 49:9 (1998) 38-52.
4. Andrew McMichael. "The Historian, the Internet, and the Web: A Reassessment," Perspectives: American Historical Association Newsletter 36:2 (February 1998) 29-32. Melanie Shell Weiss and Mark Lawrence Kornbluh. "H-Net: Humanities and Social Sciences Online," History Teacher 31:4 (1998) 533-538.
5. Dennis A. Trinkle, et. al. The History Highway: A Guide to Internet Resources (Armonk: M. E. Sharpe, 1997).
6. Joan K. Haas, Helen Willa Samuels, and Barbara Trippel Simmons. Appraising the Records of Modern Science and Technology: A Guide (Boston: Massachusetts Institute of Technology, ). This resource details difficulties with collecting records in the recent history of science and technology at corporation as a result of proprietary concerns. It also discusses early challenges posed by machine readable records and the uncertainty of how to address rapidly changing computer and software technology. Also see Clark A. Elliott, ed. Understanding Progress as Process: Documentation of the History of Post-War Science and Technology in the United States. Final Report of the Joint Committee on Archives of Science and Technology (Chicago: The Society of American Archivists, 1983) for discussion of unique factors in collecting and preserving documentation of the recent history of science and technology.
7. For an excellent discussion of intellectual property issues with regard to the Internet see National Research Council, Committee on Intellectual Property Rights and the Emerging Information Infrastructure. The Digital Dilemma: Intellectual Property in the Information Age (Washington, D.C.: National Academy Press, 2000).
8. George Glaser, William T. Coleman III, Paul N. Edwards, Henry Lowood, and Keith Uncapher. "Charles Babbage Foundation Software Task Force, Final Report," (November 1998).
9. Jeffrey R. Yost and Robert W. Seidel. "Building a Future for Software History," NSF 9979981.
10. The Charles Babbage Institute plans to continue the software history project after its three-year first phase and move on to address software development in the personal computing era.
11. Douglas R. Hofstadter. Godel, Escher, Bach: An Eternal Golden Braid (New York: Basic Books, c1979, 1999edition).
12. The Sloan Foundation funded site containing information on the development of the Boston Central Artery/Tunnel and the history of the computer mouse can be found at the following URLs respectively: http://libraries.mit.edu/rotch/artery and http://sloan.stanford.edu/MouseSite/Forum.html
13. The changing terminology as well as the fact that a number of significant publications are not indexed in the 1960s and the early 1970s makes secondary research difficult. The bibliography that CBI has created (and continually adds to) has been developed by going through the actual journals and other secondary and primary sources to overcome these challenges and provide as complete a list of available information as possible.