General Session: History and Historiography of Science Information Systems

• Overview of the History of Science Information Systems (Michael Buckland)
• Federal Funding of an Information Revolution (Thomas P. Hughes)
• Facts to Fax: Cold Fusion and the History of Science Information (Bruce W. Lewenstein)
• From 3-D Structure to Rational Drug Design: Shaping Biomedicine as an Information Science (Timothy Lenoir)
• Secret Scientific Communities: Classification and Scientific Communication in the DOE and DOD (Robert W. Seidel)

Session I: Science and Scientific Information Systems

• The Role of the Scientific Paper in Science Information Systems (Bernd Frohmann)
• The Game of the Name: Nomenclatural Instability in the History of Botanical Informatics (Geoffrey Bowker)
• Common Names: An Introduction to the History of Cooperative Access (Bernadette G. Callery)

Session II: Chemical Information Science Systems

• The Evolution of the Secondary Literature in Chemistry (Helen Schofield)
• Chemical Abstracts Service: Ninety Years of Innovation in Chemical Information (Kenneth Ostrum)
• The History of Handling Technical Information in DuPont (Florence H. Kvalnes)

An introduction to the scope, status, and prospects of studies of the history and heritage of science information systems and a discussion of historiography of this area. History is narrative of events in time past. The continuing consequences of those events are heritage, which includes our collective memory, our understanding of history. The heritage is of additional significance because the design and operation of science information systems is a professional field.

The history of science information systems overlaps the history of science and the history of information systems (previously "documentation") and, because the latter is practical, the history of technology also. There are the usual forms and genres of historical inquiry--archaeology, biography, cliometrics, and documentary research--with their differing strengths and weaknesses.

Science and technology information systems have had a privileged existence because of industrial and military needs and government policies. Much of the pioneering work concerned chemistry and/or was done by chemists. The past decade of work on the history and heritage of information systems will be summarized. Several initiatives have been undertaken to try and build a supportive infrastructure, which is important if historical research in this area is to be sustained and flourish. This conference is itself a significant part of that effort.

Since World War II, the federal government has played a major role in launching and adding momentum to an information revolution which now takes pride of place among the recent technological achievements of the United States. Federal funding not only financed the development of most of the nation's early digital computers, but financed breakthrough areas as wide-ranging as computer time--sharing, networking, artificial intelligence, and virtual reality. The government has also continued to support the education of undergraduate and graduate students who now populate industry and academic research and development centers.

Debates about the management of scientific information have traditionally assumed that information is generated first in scientific laboratories or field sites, becomes "real" or stable when it appears in peer-reviewed journals, and is then further disseminated through textbooks, encyclopedias, trade journals, government reports, mass media stories, and the like. Classic texts from the proceedings of the 1948 Royal Society Scientific Information Conference and the 1958 International Conference on Scientific Information, through modern texts like the annual proceedings of ASIS meetings, focus on how to classify and retrieve "real" scientific information that is, laboratory and field reports that have been produced for and vetted by the peer review system, as well as technical reports, patents, and other "primary" scientific information. In this paper, I will use the cold fusion saga of the late 1980s and early 1990s to suggest that communication among scientists uses many more media and much more interaction between primary and secondary information than traditionally has been assumed. This particular historical episode will suggest that we need to develop new models of the science information process, ones that account for permeable boundaries between primary and secondary information, between formal publications, preprints, electronic computer networks, fax machines, mass media presentations, and other forums for scientific discussions.

World War II saw the extension of military classification systems to a number of areas of fundamental science that had previously enjoyed unencumbered access to information. Although scientists resisted classification, and strove to publish much of their wartime work once the conflict was ended, ongoing work in fields with military applications continued to be classified, with the amount of classified information rising exponentially.

Since the high technology used by the Atomic Energy Commission (the predecessor of the Department of Energy (DOE)) and the Department of Defense (DOD) required scientific research and development on a large scale at scattered sites, technical reports, classified meetings, and even classified journals were required to support the secret scientific communities of the Cold War. In this way, the institutional values of science were replicated within those communities.

The closed world of defense science drew upon the open world of academic science and technology and, because of the participation of many private contractors, upon the proprietary world of industrial research as well. In the former case, communication was primarily one-way, although consultants from major research universities often learned a great deal from their work with the secret scientific community. In the later case, a careful screening of national security information (based on the "need to know" principle), and of proprietary information (based on the "need to sell" principle) occurred.

In order to facilitate wider exchanges of information, the Department of Defense created its Defense Technical Information Center in the 1950s, which performed services similar to the National Technical Information Service within the community of defense contractors. The Department of Energy used the Nuclear Science Abstracts to disseminate cleared information and distribution lists controlled classified documents.

In addition, areas of considerable interest in the defense community were the subjects of classified meetings, where contractors and their military sponsors could discuss common problems and present brief papers on their accomplishments. These meetings grew in number and in size as the defense R&D community waxed in the hot light of the cold war. By the 1960s, the combination of meetings, journals, technical reports, on-line information systems, and other apparatus had formed the sinews of a simulacrum of the larger scientific world within the tight skin of national security.

I will describe the history of this system, as I have learned it through probes of cold war science and technology, in the national laboratories and the Department of Defense, in order to trace the physiognomy, if not the anatomy and physiology, of Secret Scientific Communities.

The scientific journal article presents a paradox for studies of science information systems. The journal literature is considered the primary medium of scientific information, even the primary product of scientific activity. Yet studies of research front science show that journal articles do not provide the information most useful to this work. The paradox is usually resolved by appeals to establishment of priority claims and subsequent conferrals of credit through publication. Such appeals, however, ignore the content of science.

This paper argues that the paradox arises from epistemological views of science. Such views interpret science as a unified field of propositions, which are justified by scientific method, and which constitute a unified representation of nature. Scientific activity consists in formulating and testing theories, proposing and attempting to falsify hypotheses, drawing inferences, deducing consequences, and making arguments. In epistemological views of science, information plays a key role, because scientific activity is reduced to logical operations on the epistemic content of scientific statements, viz., as information processing. Thus the journal article is marginalized as a source of the required information, because its epistemic content is too old for research-front information processing.

Recent studies of scientific practices (SSP) have developed an alternative, materialist view of science, emphasizing experimental work performed in the laboratory, the primary site of knowledge production. Scientific activities are shown to consist in intervening in the world rather than in representing it (Hacking). Scientific phenomena are the products of a wide range of heterogeneous elements brought together only in particular, local, laboratory contexts. The choices, decisions, and evaluations performed in the laboratory do not conform to epistemological versions of scientific rationality, but are opportunistic, practical, strategic, craftlike, and highly situated.

Such studies have rejected the utility of a wide range of epistemological concepts in analysing science, including epistemic content, and, by implication, the concept of information. This paper argues that information seeking, information use, and information processing are of little use in understanding scientific labour. As the concept of information loses its importance in descriptions of scientific practices, questions about the role of inscriptions, utterances, graphical representations, and other material, discursive resources become more urgent. Yet few analysts of scientific practices have asked such questions.

This paper deploys several key concepts from SSP to argue that the journal article, as a particular, material discursive resource, plays a central role in scientific labour. These concepts are: the heterogeneity of the elements brought together in laboratory labour (Hacking, Latour, Knorr, Pickering); the genesis, development, and construction of facts (Fleck, Latour, Ravetz); resource standardization (Ravetz, Rouse); epistemic alignments (Rouse); epistemic things (Rheinberger); the rhetorical form of the journal article (Knorr); the role of narrative in scientific work (Rouse, Bazerman). The paper seeks to dissolve, rather than resolve, the paradox of the journal article, but without appeal to the communication of information, or to priority claims through publication.

Botanical nomenclature of vascular plants dates back to the first edition of Linneaus in 1753. Since then, there have been a series of attempts to deal with the problem of name stability: how to ensure that a given species will have the same name all over the world and over time. In the 1820s the Kew Rule was developed out of Kew Gardens to deal with priority in generic names; in the 1860s George Bentham worked on two large projects to stabilize systematics. In the 1890s the Berlin rule (limiting priority for names which had fallen into disuse or never been accepted) came into conflict with the Philadelphia rule (according to which priority was absolute.) Throughout the past century there have been a series of international conferences to deal with issues arising.

The principle issues are twofold. On the one hand, it is highly desirable to be able to change the names of plants when new scientific insights come into place. On the other hand, it is extremely difficult and costly to change the names of plants. In this paper, I explore the developing positions with respect to name stability. First, with respect to the nature of scientific cooperation (here in the form of the establishment of international conferences to deal with questions arising, and the development of agreements about which scientific journals could carry new names). Secondly, with respect to the development of information technology (computer technology, for example, makes name flexibility in some ways easier to propagate). I argue that the issue of name stability has been the site of a series of significant discussions about the nature and storage of scientific information. This constitutes a development of recent work of mine about the history of classification systems in medicine (Bowker and Star, forthcoming) where I have argued that work done at the deep infrastructural level of classification and nomenclature systems is closely tied with at once information technology developments and organizational histories (the organization of the profession of medicine or botany in these cases). I also argue in this paper that it is impossible not to encode some specific readings of the state of knowledge and of the state of relations between often competing professional groups (sytematists, botanists, farmers) deeply into the information infrastructure: and that this has had significant consequences both for knowledge production and organizational change.

References:

Geoffrey C. Bowker and Susan Leigh Star, How classifications work, Cambridge, MA: MIT Press, forthcoming 1998.

D.L. Hawksworth, Improving the Stability of Names: Needs and Options; Proceedings of an International Symposium, Kew, 20-23 February, 1991, Konigstein, Germany: Koeltz Scientific Books, 1991

The design and use of electronic information systems to provide cooperative access to natural history museum collections is influenced by existing traditions of organizing paper-based information about those collections. These information systems are characterized by a focus on the uniqueness of the collected specimen and a lack of standardization of the choice of identifying characteristics and the descriptive terms to specify those characteristics. In addition, the scientific name initially assigned to a given specimen may change, either due to incomplete or incorrect initial identification or subsequent changes in the nomenclatural hierarchy of the organism. While these inconsistencies could be accommodated by a limited staff in an individual museum, such information systems may not serve the needs of a more diverse group of users, especially if both academic and public users are included. A discussion of historical museum recordkeeping systems provides background for a discussion of current museum information systems designed to provide cooperative electronic access to multiiple collections. While the rhetoric of network culture may imply that the technology that enables cooperation will ensure that cooperation, such cooperation is not easily achieved. Information management techniques drawn from the library and art world for providing electronic access to records with dissimilar form, content, and descriptive vocabularies are offered as possible alternative models.

Top of the page

Information scientists and librarians commonly define secondary publications as tools which facilitate identification of relevant primary publications. In this paper, the early history of secondary publications in general is considered briefly, commencing with the first abstracting journal, the Journal des SIavans in 1665. The paper then concentrates on the evolution of the secondary literature in chemistry. Chemistry was one of the first branches of knowledge for which specialist secondary sources were published, with the first such publication being Crell's Chemisches Journal fur die Freunde der Naturlehre in 1778. The historical reasons for the need for secondary sources in chemistry are examined, which include the increase in the number of chemistry journals in existence and developments in the discipline of chemistry. The historical background to a number of the most important secondary sources in chemistry are given, including in some cases short biographies of their originators and producers. Some of the most important sources, including Chemical Abstracts (which commenced publication in 1907), Beilstein Handbuch der organischen Chemie (1881) and Gmelin Handbuch der anorganischen Chemie (1817) are discussed in some detail. The differences in editorial policy and criteria for inclusion of publications amongst the secondary sources are discussed. The organisation of chemical information in several secondary sources is presented along with how classification schemes were developed to aid retrieval of information. Variations in the arrangement of information in several publications are identified, which can be due to the differences in the nature of the information included in the publication, differences in editorial policy and differences in the required access points to the information. The majority of chemical information requests are based on the need for information about chemical compounds, which in turn centre around the two- or three-dimensional chemical structure. The special nature of chemical information and the language of chemistry in this context is discussed. The evolution of index systems in the pre-computer era which make use of chemical names and formulae to enable access of chemical information is described, including their ability to deal with new types of substance as they have been discovered. An appendix which lists major secondary sources in chemistry, both those still published and those which have ceased publication, is given.

Part of the chartered purpose of the American Chemical Society (ACS) is "increasing and diffusing chemical knowledge" for the benefit of industry and education. The printed publication CHEMICAL ABSTRACTS (CA) began in 1907 as a member service of the ACS to help scientists identify relevant information from among the thousands of scholarly studies and patents available. The goal for Chemical Abstracts Service (CAS) today - as a self-supporting division of the ACS - is the same. Only the amount of information available and the tools used to scrutinize it have changed. This paper will summarize the many electronic distribution channels used by CAS today to make chemical information available to scientists around the world in a timely manner with the quality and comprehensive coverage that have been the hallmark of CAS since 1907.

Traditionally, DuPont has supported three areas of technical information: library services, patent information and proprietary technical information. Library services and patent information handling will be discussed briefly. The handling of technical proprietary information will be discussed in greater detail.

In the 1960's, Dupont developed computer databases to manage its collection of technical information. The handling of information was affected by the organization of the DuPont company. Many of the individual departments had their own information centers.

In the 1960's, studies were undertaken to determine the duplication of effort between departments, especially in the area of patent services. These studies also impacted the handling of proprietary technical information. The development, building and implementation of an integrated system for the storage, retrieval and distribution of DuPont scientific and technical information will be reviewed.

In addition to changes in the mechanical handling of information, changes were made to the way the content was described, eg from classification systems to concept coordination. Many of the fundamental principles developed for storage and retrieval of technical information developed in the 1960's were adopted and are in use today. Innovations in software and computer technology have enhanced the implementation of the potential and capabilities of these earlier principles. The culmination of these innovations resulted in the SCION database created in the early 1990's. This database provides online access for the DuPont scientific community to their proprietary technical information.

Before then, there had been several batch systems. In the beginning, they were independent systems which evolved by collaboration to a common system used by the still independent departmental information groups. In 1964, the Central Report Index was formed by the combination of the information groups of nine departments. An important development for the integration of the nine separate systems was the combination of nine different thesaurii into one hierarchical thesaurus.

During the same time period, two DuPont engineers developed an algorithm which permitted the storage and retrieval of chemical information by structure or atom and bond connection tables. Subsequently, this information was shared with Chemical Abstract Service and served as the foundation for the development of the Registry File. Concurrently, the Central Report Index began development on a computer system that would fully integrate the files of the previous nine independent indexes and provide searching of chemical structures as well.

These earlier computer systems were batch systems accessible to only the staff of the Central Report Index. With the advent of online searching of external literature through Dialog, Orbit, Chemical Abstracts Service, not only the DuPont scientists, but the information professionals in the Central Report Index and other DuPont information centers demanded online access to the proprietary information in the Central Report Index's batch database. As an interim measure, in the mid-1980's, a bibliographic and abstracts only database was developed that was available online through DuPont telecommunications networks.

The next generation of the SCION database is under development and will take advantage of the new generation of computer and information science techologies, eg web browsers, natural language searching.

Top of the page