Charge

This task force will consist of university administrators, librarians, and scientists that collectively bring expertise in scientific communication, library and information management, and telecommunications networks. It will examine new options for the collection and dissemination of scientific and technical information, giving particular attention to the concept of a distributed national science and technology library, a system of regional libraries responsible for the collection and dissemination of scientific and technological journals and related forms of communication.

Background: The rapid development of communication and computing technology is changing the way scientific information is created, disseminated, and used. Nonetheless, the preponderance of scientific information is provided through serial publications that are rapidly growing in number and price and are increasingly concentrated among a small number of commercial publishers.

The economic pressures on a captive university market, combined with new technological possibilities for handling information, can provide the motivation and the means to develop a system of regional libraries that would provide a nationally organized structure for managing scientific and technological information. Among other functions, such a system could collect and store serial publications which could be distributed electronically to local users. Canada has created the Canada Institute for Scientific and Technical Information which supports the highest level of interlibrary lending in North America. A similar system, regionally distributed to develop broad political acceptance and to accommodate the greater dimensions of the U.S. scientific and technological information enterprise, could provide a structure that would reduce collection requirements--and therefore costs--for individual universities while sustaining and even expanding access by students and faculty to scientific and technological information.

The computing networks currently being developed by the federal government and the private sector will provide the electronic communication system for handling national science and technology information flow. It is important for universities to participate in the development of such a system so that it reflects both the research and educational needs of those institutions.

I - OVERVIEW AND RECOMMENDATIONS

A. Task Force Charge

The work of this Task Force is part of the Association of American Universities Research Libraries Project, undertaken in collaboration with the Association of Research Libraries with the support of The Andrew W. Mellon Foundation. Since the initial project meeting in January 1993, some three dozen individuals from the academic community - librarians, faculty, technologists, and academic administrators - have worked in three Task Forces to focus university-wide attention on issues confronting university libraries and to identify long term, economic solutions for effective scholarly communication. The members of this Task Force were asked to address issues surrounding scientific and technical information (STI).

The Task Force on a National Strategy for Managing Scientific and Technological Information was charged by AAU to examine new options for the collection and dissemination of scientific and technical information that could break an extraordinary information cost spiral while at the same time support innovative applications of information technology. Specifically, the Task Force was asked to address the potential of a distributed national science and technology library, formed by a system of libraries to coordinate the collection and dissemination of STI journals and related forms of communication.

B. Advantages of New Models for Managing STI

To evaluate different options for managing STI, the Task Force adopted an analytical framework that describes the system of scientific and scholarly communication. The result of an analysis of that framework was three distinct models that depict different approaches to information management for STI -- the Classical, Modernized, and Emergent models (described in Section II and elaborated in Appendices B and C).

The term "model" is not used in this report in a predictive or prescriptive sense, rather the models are descriptive representations of ways that scholarly and scientific communication operate now and could operate in the future.

The Task force confirms that the technologies at hand -- and their creative application -- present opportunities for research libraries and universities to adopt new strategies to improve current library services as well as to make a qualitatively new contribution to the management of STI. Even so, institutions will be able to benefit from the advantages presented by new approaches to managing STI only when library services and operations are reengineered to support the new management models and when campus constituencies are provided with network-based, desktop access to remote STI resources that is convenient and reliable. At present, the Modernized and Emergent models tend to overlie and supplement rather than replace the Classical model, and while they offer advantages, the full scale adoption of these approaches is neither inevitable or easily accomplished.

C. Potential for Innovation

New technological tools for research, electronic networks, networked information, and creative application of information technologies are already transforming communication among scientists, scientific publication, and the way scientists receive and use information and data. The Task Force approached its assignment therefore with a shared vision of scientists with desktop access to campus, national, and even global networks and services that identify, locate, and deliver multiple forms of scientific and technical information, regardless of the physical location of that information.

Through the ability to express all media digitally, it is technologically feasible now to provide scientists with integrated access to STI collections in multiple forms. For example, the journal article and other page-formatted documents may be linked with scientific data, images and sound collections, as well as other forms of STI research resources. Hyper-media links and extends a page image to related information resources and will increasingly serve as an interface to the environment of networked information. Early forms of these capabilities are now in growing use through the rapid deployment of WorldWideWeb servers accessed through Mosaic. It is within such a dynamic environment that "digital libraries" or large collections of electronic resources are actively being defined and created.

Simultaneously, a new kind of scientific communication process has emerged that takes place entirely or primarily on electronic networks. For some applications, in particular those that involve teams of collaborating scientists, this process represents a fundamentally new approach to scientific inquiry. Traditional patterns of sharing research results are altered or bypassed. "Publication," as the academic community has come to recognize it, expands, and new ways are used to disseminate the dynamic research results for the use of other scientists now and in the future.

D. Potential for Cost Containment

For many participants engaged in the process of scientific communication, the costs of managing STI are hidden. This is largely due to the complexity of the process and the number of participants in the system. For most university scientists, STI acquisition costs are borne indirectly, in another part of the institution's budget -- the library. In fact, to address the problem of rising costs for STI research resources raises issues that extend beyond the research library and which are embedded in the economics of scientific publishing and the very culture of the university. Long term solutions are, therefore, complex.

Cost containment for collecting and disseminating STI within a university depends upon factors such as the following:

• The extent to which academic authors of STI retain copyright ownership and control of their intellectual property in a not-for-profit publishing environment;

• The extent to which universities and their libraries are able to influence the prices charged for acquiring the results of scientific research;

• The extent to which universities provide the tools and competitive outlets to receive, distribute, and archive STI; and

• The extent to which new strategies for managing STI minimize the need for every research library to maintain comprehensive, paper-based collections of STI.

Use of electronic networks and electronic publishing can perceptibly lower the costs associated with publishing (by eliminating cost for paper, printing, and distribution). However, if the educational and research community is to benefit from the lower unit costs from the use of electronic technologies, more competition must be injected into the current monopoly-like marketplace in order to moderate subscription prices.

The Impact of Copyright and a Monopoly-like Marketplace on Prices for STI

Most sciences are now served by an imperfect, monopoly-like marketplace for STI publishing. In the recent history of scientific communication, the publication of many STI journals has migrated into the hands of a few, very large commercial publishers. In a commercial environment of this nature, it is unrealistic to expect significant reductions in the cost of a library's materials budget for STI.

Scientists seek to publish in the most prestigious journals of their field. A common practice in the system of scientific publication is for the scientist author to be required to assign copyright to the publisher as a condition of publication. The institution then repurchases the article in journal subscriptions, using funds from library and/or departmental budgets. Because of the "scarce good" nature of the scientific journal -- in particular, the prestigious titles in each discipline -- there is no market substitute; publishers can and do charge extraordinarily high prices for these subscriptions. Because the ownership of articles is assigned to the publisher, the profit returns to the publisher -- not to the scientist author or to the institution that may have supported the research. And the control for the duration of copyright belongs to the publisher.

This situation has an unfavorable impact on the educational and research community because it allows prices for STI journals to be raised at rates that far exceed inflation. Until circumstances change (until the market stops buying), the cost of a library's materials budget for STI, especially for journal subscriptions, will continue to increase. In addition, it is reasonable to expect that the copyright fee imposed for the use of individual articles (over and above those uses granted by U.S. copyright law for educational and fair use) will be set by the publisher at levels that would make up for any revenue lost from cancellation of journal subscriptions. Until the marketplace for STI is changed in ways that cause STI publishers to lower their revenue targets, libraries and their universities should expect materials budgets for STI journals to continue to rise, whether the material is purchased as a subscription to a journal or singly from an article delivery service.

Therefore a key university strategy to realize long term cost advantages in the acquisition of STI is to introduce more diversified competition into the marketplace and thus moderate the prices charged for science journals and articles. The development of cost-based (instead of market-priced) electronic STI publications with high editorial standards could interject this competition into the current environment. University reward and recognition systems must also embrace electronic publishing as a legitimate outlet for publication of research results. Electronic journals must be able to compete in terms of academic legitimacy and editorial quality with traditional paper journals.

Electronic publishing of the results of scientific research is not an insignificant burden to ask universities or other not-for-profit organizations to assume. In some cases, the Task Force recognizes that additional financial support for such organizations is required to make this concept a reality. However, creation of competitive outlets for publishing STI remains the single most important action with potential for a long term solution to contain the costs associated with managing STI.

Pressures to Maintain Paper Collections of STI

The role of the library is to identify significant STI collections (in all formats); make judgments about which of these resources to acquire, organize, catalog, and index; provide current users with expert, professional services to make full use of the collections; and assure future users of access to significant collections through library preservation and maintenance of comprehensive research-level collections.

New information management strategies that exploit networks and new technologies present opportunities for improved performance of these functions, in particular the capacity to deliver STI to the desktops of scientists and students. These approaches also offer the potential, through resource sharing strategies such as a distributed national library for STI, to eliminate redundancy and therefore contain costs for some local library management functions such as cataloging, storage, and preservation. (However, under existing marketplace conditions for STI journals this approach alone should not be expected to result in long term savings in library materials budgets for STI.)

Factors affecting full-scale library adoption of new information management strategies include campus readiness to substitute desktop access for locally maintained collections of STI, and/or the extent to which the new publication and management options are developed to perform all the functions required for effective support of scientific communication. For example, until electronic articles or journals are archived to ensure their integrity and preservation for future users, scientists will continue to expect the library to acquire, maintain, and preserve journals in paper. Similarly, until a campus is networked (including "the last 100 feet") and access to desktop service is ubiquitous and reliable, convenient access to paper collections will continue to be a priority for scientists and students.

E. The Research Library and Campus Readiness for Change

Libraries have an impressive record of using technology and electronic networks to improve access to research collections and to automate bibliographic operations such as acquisitions and cataloging. In addition, in the last several years there has been an increase in library use of computer networks to send and receive page images of articles or other kinds of STI research materials. Libraries are also engaged in the establishment of standards and mechanisms to describe and locate electronic resources. The availability of digital full text on the networks may change the mechanisms used to describe and locate resources, but the function remains a critical element of the scientific communication process.

Campus information systems and policies need to accommodate the demands of a transformed system of scientific communication. It is also essential to ensure that any new national networking programs build upon and extend the progress achieved to date by NSFNET and provide universities with reliable network connectivity, with ease of use, and at a reasonable cost.

In addition, a campus-wide readiness is required if information professionals are to take full advantage of new technologies to deliver library services directly to scientists and students and to expand the range of STI resources made available via networks. This is in part technical readiness (networking, hardware, software, training); in part cultural readiness (change in traditional university practices and values); and in part leadership readiness to convince campus constituencies that a transformation in the institutional management of STI will improve access, will support the creation and management of knowledge for all disciplines, and will significantly strengthen the capacity of the university to pursue its mission.

It is important that science faculty, as authors and as consumers of STI research resources, collaborate with librarians to shape new options for managing STI. It is also important for university leaders to provide an impetus to secure necessary change in current university practices that will benefit scientific research and the academic community as a whole.

The changes the Task Force identified as necessary are not ones that can be mandated. Instead, they must be pursued by encouragement, example, and incentives. The recommendations presented do not offer simple solutions to "fix" the STI problem. In an environment of multiple participants, fast-changing technology, and limited resources, each institution needs to pursue a long term strategy for managing STI. What the recommendations propose are specific steps that AAU and ARL could take to help individual institutions to develop and implement such a long term strategy.

F. Task Force Recommendations

1. Ensure that electronic networks and networking policies are in place to take full advantage of the technologies at hand and accommodate the demands of a transformed system of scientific communication.

To ensure that electronic networks and networking policies are in place to take full advantage of the technologies at hand and accommodate the demands of a transformed system of scientific communication, AAU, in collaboration with ARL, should assist their member institutions to pursue three kinds of actions:

• Collectively, exert an influence within national, state, and provincial governments to ensure that the legal and regulatory framework for communications will support affordable higher education access to a structure of national and global information networks (such as the information superhighway envisioned by the current U.S. Administration);

• Collectively, promote public and private funding sources to assist higher education in creating and preserving digital collections of STI, and to encourage innovative network applications; and

• Individually, establish a process to determine how to speed the completion and support of reliable network access to the offices of scientists and others who use STI, and to the dormitories and libraries used by students of the sciences.

2. Introduce more competition and cost-based pricing into the marketplace for STI by encouraging a mix of commercial and not-for-profit organizations to engage in electronic publication of the results of scientific research.

AAU and ARL should each adopt a policy statement that calls for more competition in the marketplace for STI publishing and encourage not-for-profit organizations to engage in cost- based electronic publication of STI research.

AAU, in collaboration with ARL, should provide their members with information on how a university can encourage not-for-profit organizations to create electronic journals that can achieve high editorial quality and prestige in the scientific community. Examples might be illustrations of institutional leadership, financial support, and university partnerships with scientific societies to establish, operate, and archive peer reviewed electronic journals, as well as examples of incentives that redirect copyright assignment for STI intellectual property from the commercial to not-for-profit publishers, or policies that support quality electronic journals as legitimate publishing outlets for faculty.

Comparable information could be exchanged with the Higher Education Funding Council of England, Committee of Vice Chancellors and Principals, a United Kingdom organization that is also seeking practical and effective ways to influence the marketplace for STI journals. In late 1993, the UK Committee received a commission report addressing the future of libraries in support of UK education and research in which issues comparable to those addressed in this report were raised. AAU was specifically identified as a potential trans-Atlantic partner in addressing the periodical pricing problem.

3. Explore the feasibility of actions to mandate retention of the ownership of certain STI intellectual property in the not-for-profit sector.

AAU, in collaboration with ARL, should commission one or more inquiries to assess the feasibility of the following actions to mandate retention of ownership of certain STI intellectual property in the not-for-profit sector.

• University action: University policies, practices, and mechanisms should retain in the not-for-profit sector some or all of the ownership rights of STI intellectual property that is created by university faculty and staff.

• Government action: Agencies of the U.S. Government that fund scientific research should, as a condition of funding, require that a copy of any report, data sets, or articles developed as a result of publicly-funded work, be deposited into an electronic repository that will provide public access.

The results of inquiries into the feasibility of such actions should influence an AAU implementation strategy to encourage retention of STI intellectual property in the not-for-profit sector.

For options for university action, see the Report of the AAU Task Force on Intellectual Property Rights in an Electronic Environment. For a proposal for action by U.S. Government science agencies, see Section V of this report.

4. Promote establishment of a system of national repositories for scientific research to establish not-for-profit electronic outlets for STI .

AAU, in collaboration with ARL, should take the lead to convene a meeting to develop an implementation plan for the proposed system of national repositories for STI as outlined in section V of this report. This system of distributed electronic repositories would provide the organizational structure needed to manage digital libraries of STI.

5. Undertake a demonstration project to test the concept of a distributed national science and technology library.

AAU, in collaboration with ARL, should test the concept of a distributed national STI library (see Section III B of this report) through a demonstration project with a consortium or federation of institutions with strong library collections of STI. Unless the marketplace for STI is changed, a distributed STI library will not lead to long term savings in library materials budgets but it can improve access and minimize the need for every research library to maintain paper-based collections of STI.

The proposed demonstration project centers on STI resources on biotechnology and computer sciences from Japan. This project was identified by the AAU Task Force on the Acquisition and Distribution of Foreign Language and Area Studies Materials.

6. Invest in and evaluate the results of new ways of managing STI (particularly ways that realize the Emergent model).

AAU, in collaboration with ARL, should monitor experiments or ongoing demonstration projects that seek to advance Modernized or Emergent models for managing STI, and produce a periodic assessment of these efforts in the context of the Task Force findings. Every such project should be assessed to determine the extent to which all the functional requirements of the system of scientific communication are performed, highlighting their acceptance by scientists and students and their cost to the institution.

7. Promote the awareness of the Task Force findings among all participants in the scientific communication process.

AAU, in collaboration with ARL, should encourage and support understanding and discussion of these findings within universities, in organizations of faculty, librarians, or university leadership, in scientific society forums, with agencies of the U.S. Government having responsibilities for science or STI, and with current and potential electronic publishers of scientific and scholarly information.

A heightened awareness among all participants in the scientific communication process, especially among science faculty in research universities, is essential to highlight these issues and the consequences of perpetuating the status quo.

II - The System of Scientific Communication: A Framework for STI Management

To evaluate different options for managing STI, the Task Force adopted an analytical framework that describes the system of scientific and scholarly communication. The framework identifies all functions of that system (e.g. authoring, informal peer communication, editorial review and validation, acquisition and access, location and delivery, preservation and archiving); the major participants in that system; and the performance attributes by which such systems are judged (e.g. timeliness, authenticity, cost).

The elements of the scientific and scholarly communication system as defined within the framework are elaborated in Appendix B of this report.

The framework provides a basis for comparing and contrasting different ways that scientific and technical information is created, shared with others, and used. The description of the framework may suggest a linear, orderly progression of scientific communication, even though the process itself is anything but. The functions may be viewed as "modules" which are performed at various times (or simultaneously) and which can be re-arranged in different information management scenarios. The framework accommodates all information formats and multimedia information. What will change, depending on how the information is managed, is the sequence in which the functions are performed, how well they are performed, and which system participants have responsibility for different functions at different points in the overall process.

Using the components of the communication system identified in the analytical framework, the task force formulated and evaluated different models of information creation, dissemination, and use. The term "model" when applied in this report to information resource management (IRM) is not used in a predictive or prescriptive sense. The models the Task Force identified are intended as descriptive representations of various ways scholarly and scientific communication operate now and could operate in the future.

A. Three Approaches: Classical, Modernized, and Emergent

The result of the framework analysis was three distinct models that depict different approaches to information management for scientific information. The STI management models are characterized as Classical (print-based), Modernized (defined in three tiers with a mix of paper and electronic inputs and outputs and increasingly direct computer access to university-generated writings), and Emergent (in which scientific collaboration, communication, and information sharing takes place entirely, or predominantly, on an electronic network).

The three scientific and technical information management models are described briefly on the following page. Figure 1 displays the key characteristics of the information transfer flows and how they differ in each of the three models; the models are described in greater detail in Appendix C.

Classical: The Classical, print-on-paper system of STI is exemplified by the traditional scientific journal. The print journal remains, for most fields, a primary model of scientific and technological communication. Traditional institutional support for managing STI was predicated on this model. For example, libraries were established and funded to buy, house, preserve, and provide retrieval services for STI. The rising cost of maintaining this approach to managing STI is a primary impetus to the work of this Task Force.

Figure 1

Modernized: The Modernized model is to a great extent the current model, in that most of its functions are available today and overlie the Classical model. In the Modernized model -- which could also be called the "article delivery," "on demand," or "just in time" model -- information is still largely published and resides primarily in print form, but there is an array of expanding options for discovery and delivery. (An example of the transition from the Classical to the Modernized model is the evolving system for handling interlibrary loan and document delivery transactions; see Box 1.)

The Modernized model can also accommodate electronic journals that are authored and published in electronic forms and made available to library users via electronic channels. In the Modernized model, libraries can acquire fewer resources for local collections and rely on a growing number of remote sources for retrieval and delivery. An article or book need be acquired only if and when the user requests it (in paper or electronic form). The application of the Modernized model in this way presumes, however, that comprehensive collections of STI resources are still developed and preserved somewhere in order to supply libraries with this article delivery service. The issue is how to define the circumstances when it is advantageous to treat a given library as a consumer or as a provider of STI article delivery services.

Emergent: The Emergent model uses computing and communications technologies to share scientific instruments, primary data, and software as well as to support scientists in communicating and sharing information per se. These technologies are used for genuinely innovative purposes, rather than for the modernization of page-formatted information from paper to network storage and access media. Information environments generated by this model are sometimes referred to as "collaboratories." Examples include the human genome project and the global climate change initiative (briefly described in The Landscape of STI, contained in Appendix A to this report). Box 2 also gives further details about the Emergent model.

The Classical and Modernized models share the characteristic of information management as "post publication" functions. In each, the librarian or information manager typically becomes involved in the scientific communication process after publication of an article, journal, or book. The Emergent model differs in this respect, as it assumes involvement of librarians and other information managers earlier in the process of scientific communication. This distinctive feature of the Emergent model has implications for changing roles and responsibilities within institutions.

B. Why Universities Should Expect a Mix

Universities should expect that a mix of scientific and technical information management models, including the Classical, is inevitable. Figure 2 postulates the mix of management models for the next twenty years. The deployment of multiple models of information management will not progress uniformly largely because not all scientific disciplines will embrace electronic technologies at the same pace. In as much as the rate of institutional change will vary, it is realistic to assume a dynamic mix of IRM models for several decades.

The Task Force's review of current and anticipated applications of the models for STI concludes that the Modernized and Emergent models tend to overlie and supplement rather than replace the Classical model. This is due in large measure because applications of the Modernized and Emergent models are not yet fully developed. The transition from the Classical to the Modernized and Emergent models for STI is not inevitable.

It is important for an effective information management strategy to perform all of the functions of the system of scientific communication. For example, for electronic journals or articles to substitute for their paper equivalents, the electronic versions must receive peer review, be accepted as equal in prestige to traditional paper journals, and be archived to ensure their integrity and preservation for the future. While the task force reached consensus that the Classical, paper-based model is, over the long term, both unaffordable and unresponsive to new scientific communication patterns, it concluded that it will be sustained until the other models are more fully developed.

C. Advantages of the Modernized and Emergent Models

To differing degrees, both the Modernized and Emergent models rely on networked information and/or electronic networks for dissemination of research resources. The use of electronic formats instead of paper permits improved performance, particularly in the functions of storage, access, and delivery, and potentially in acquisitions and information management (cataloging). Access and delivery to local and remote resources could be faster and funds could be reallocated from ownership of a book, journal, or datafile to payment for use of such resources from remote sources; this would reduce local costs of acquiring and cataloging material.

To the extent that the Classical model is replaced by the Modernized and Emergent models, the kind of library facilities required would also change. There would be less need for expanding stacks to shelve paper materials and more need for space configured for user workstations, for meetings among users, and for people to work together with librarians on research design and effective use of local and remote networked collections. However, until a campus is networked (including "the last 100 feet") and access to desktop service is ubiquitous and reliable, convenient access to paper collections will continue to be a priority for scientists and students, and the Classical model will be perpetuated.

It is therefore advantageous to complete the development of the Modernized and Emergent models to perform the full range of functions required for scientific communication so that research libraries can employ them and minimize the use of the Classical, paper-based model for managing STI resources. There is evidence from institutions represented on this Task Force that this shift has already begun as library expenditure trends show a slowing in the amount spent to purchase printed materials and an increase in that spent on electronic resources; institutional expenditures for networks are also going up along with usage of the networks.

Any reduction in spending on one function or activity may not result in cost savings to an institution. Some cost savings may lead to reallocations within a library budget (that can seed projects to pave the way toward the Modernized and Emergent models), and/or result in cost shifts within an institution (e.g., among the library, computer center, university press), among institutions, and/or from the institution to the individual user. The economic impact of new models of STI management is not fully predictable. What is needed are ongoing assessments of experiments and demonstration projects to inform better our judgment about where and how real costs may be contained. Examples of previous and ongoing assessments are noted in The Landscape of STI included in Appendix A of this report.

Figure 2 - STI Information Resource Management Models
Estimated Institutional Use of Information Models by Year

III - Re-engineering Library and Campus Information Services

As Figure 2 suggests, libraries will not be able to benefit from the cost advantages presented by new approaches to managing STI until the Modernized and Emergent models are developed to a point where they replace the Classical model. To speed and anticipate this transition, library services and operations need to be reengineered to support the new management models and campus constituencies need to be provided with network-based, desktop access to remote STI resources that is convenient and reliable.

A. Anticipating the Emergent Model: Building Partnerships

The information management roles or responsibilities of scientists and librarians within an institution will differ depending on the information management model. The Classical and Modernized models will continue the traditional roles and relationships. The Emergent model, however, transforms how scientists and librarians have customarily interacted and accomplished their work. The resulting new roles require organizational flexibility and may lead to structural change in an institution.

One anticipated change will be the involvement of librarians at points early in the scientific inquiry and communication cycle. In the Emergent model, teams of scientists and librarians work together to create specialized databases and tools in an interactive, electronic environment focused on the specific needs and problems of a given user community. Partnership among scientists and librarians constructs a process that maximizes the usefulness of data gathered and information generated.

The Task Force discussions did not prescribe specific ways the partnership between librarians and scientists should evolve. However, they did underscore the importance for institutions to engage in projects that build working partnerships among scientists, librarians, and other information professionals on university campuses. Such projects will combine expertise, create technology that better meets scientific needs, enable conceptualization of new research problems because tools and instruments are in place or can be developed, and ensure that information is organized early in a project. Demonstration projects that build these partnerships and capabilities will leverage the benefits of new technologies and develop new modes of scientific inquiry.

B. Anticipating the Modernized Model: A Distributed National Library for STI

The concept of a distributed national STI library presumes that certain designated libraries operate in a nationally organized structure to manage STI and provide electronic distribution to remote users, thereby reducing collection requirements -- and costs -- for individual universities while sustaining and even expanding access to STI by students and faculty. This could contain library costs for managing STI by reducing time and resources required to acquire, catalog, house, and preserve local collections. Technology makes a coordinated collections strategy feasible, and it expedites article delivery services. However, under existing marketplace conditions for STI journals, this approach should not be expected to result in long term savings in library materials budgets for STI.

The concept of a national STI library is not untested. In 1974, the National Research Council (NRC) of Canada, an agency of the Canadian federal government, established CISTI, the Canada Institute for Scientific and Technical Information. It is the successor to the National Science Library that was originally established in 1924. CISTI maintains the largest collection of published information in science, technology, and medicine in North America. The collection consists of half a million books and conference proceedings, over 50,000 serials and more than two million technical reports. The collection catalog can be accessed electronically, as can CISTI's online article ordering and electronic delivery capability. CISTI also provides online database services through its CAN/OLE (Online Enquiry) and CAN/SND (scientific and numeric data) systems. CISTI's strategic plan reflects its commitment to develop a virtual library by introducing electronic-based services as products and services become available. In collaboration with other Council research laboratories and university partners, CISTI also undertakes research and development into innovative information systems. In 1994, the NRC STI publishing program comprising 14 peer-reviewed journals is being merged with CISTI.

The Task Force was aware that a similar agency, the National Periodical Center, was proposed but rejected in the United States at about the same time CISTI was established. [6] The acceptance of CISTI in Canada might be attributed to a perception when it was established that it would expand the universe of sources available and supplement existing library collections of STI.

To revisit the establishment of such a service within the United States, the task force considered that a number of strong collections of STI resources already exist in research libraries. The establishment of an inter-institutional agreement to provide large scale article delivery services for STI should involve libraries with strong STI collections and build upon the architecture of collaboration that already exists. Participating libraries could link together existing but geographically dispersed collections of strength to serve as a "virtual national collection" of STI resources. To function as a national, North American, or global system, the designated collections would have to be maintained in coordination with each other, be network accessible and deliverable, perform a full range of quality services, and contain their costs to keep prices charged for the services at a reasonable level.

Transforming the current system of resource sharing among independent research libraries into a national, North American, or global federation raises issues of institution-wide significance. Such a multi-institutional network of coordinated collection building will require consensus and contractual agreements among research libraries and their constituent scientific communities about certain practical matters. Which institutions, for example, will commit to maintain a national STI collection over time even if the local academic or research program changes? What is the level of service performance expected of a library that maintains a national collection? How will universities allocate resources to purchase services from such national collection(s) of STI.

To extend the concept of a distributed national collection of STI collections to one that substitutes for local resources instead of expanding them raises an additional set of issues. "Consumer libraries" need to address institutional concerns about what users require in order for remote services to function as an effective substitute for comparable local collections and services. Agreements need to be struck between "consumer libraries" and "provider libraries." These agreements need to assure users that fund reallocations will enhance, not diminish resources available to them. Such agreements must also assure users that network access will function as an effective substitute for local ownership. Similarly, "provider libraries" have to address institutional concerns about the impact of this additional responsibility on service to their own institutions.

Fundamentally, consumer and provider libraries and their institutions will have to craft agreements and provide support to ensure that such a federation will enhance access to STI resources. Universities would also have to be willing to embrace new assessments of institutional quality, substituting new measures for the traditional statistics reflecting size of library collections.

This inter-institutional strategy mirrors the distributed network-based program proposed by the AAU Task Force on the Acquisition and Distribution of Foreign Language and Area Studies Materials to facilitate access to publications from overseas. A similar program for STI would significantly expand access to research resources for any scientist with network access, regardless of location.

One demonstration project proposed by the AAU Task Force on the Acquisition and Distribution of Foreign Language and Area Studies Materials would develop and evaluate network access to distributed collections of STI research resources from Japan. This is the kind of demonstration project that can test the capabilities of the Modernized model, identify the characteristics that foster the Classical model, and identify the degree of centralized management required for effective operation of a distributed system. This Task Force supports this proposed demonstration project for Japanese STI resources and urges that any assessment of its progress be undertaken in the context of the findings of this report. In particular, it should specifically track how intellectual property ownership for Japanese STI affects the cost of the materials budget during the demonstration.

Whether such a distributed national library for STI could, in the long term, contain the cost of materials budgets for STI will be determined by the impact universities can have on the STI publishing marketplace, in particular the prices charged for acquiring STI research, as discussed below.

IV - Introducing Competition to the Marketplace for STI Publishing

The single most important impediment to realizing any long term cost advantages from the new models of scholarly communication is the monopoly-like nature of STI journal publishing. There is ample evidence that scientific and technical journal pricing does not reflect a competitive economic marketplace. (Selected published literature on the topic is listed in The Landscape of STI and The Economics of Scholarly Communication: A Working Bibliography, contained in the appendices to this report.)

In the recent history of scientific communication, the functions performed by publishers (editorial, validation, and distribution) have increasingly migrated into the hands of a relatively few, very large commercial publishers. In 1989, the Association of Research Libraries released a report documenting this trend and its impact on pricing and on the ability of research libraries to build collections of journals and other serials. The report recommends the introduction of greater competition to the commercial publishers. [7] The AAU STI Task Force concurs with this recommendation.

Scientists seek to publish in the most prestigious journals of their field. A common practice in the system of scientific publication is for the publisher to require the scientist author to sign over copyright as a condition of publication. The institution then repurchases the article in journal subscriptions, using funds from library and/or departmental budgets. Because of the "scarce good" nature of the scientific journal -- in particular, the prestigious titles in each discipline -- there is no market substitute; publishers can and do charge extraordinarily high prices for these subscriptions, and the increase has been in considerable excess of inflation. As the copyright is assigned to the publisher the profit returns to the publisher, not to the scientist author or to the institution that may have supported the research. And the control for the duration of copyright belongs to the publisher.

In the last twenty years, librarians have allocated more funds to publishers of STI journals to cover price increases, not to acquire additional titles. Figure 3, a chart of monograph and serial costs in ARL libraries, illustrates that increasing expenditures for serials have resulted in fewer books and journals being acquired. In AAU universities, it is not uncommon for a research library to spend two-thirds of its serials budget for science journals -- journals that are priced largely in a monopoly-like marketplace. All AAU universities would benefit from a more competitive marketplace for STI publications. That is a marketplace where prices for journals are determined more by the actual cost of production instead of by a monopoly-like marketplace. To create such a marketplace, there is a need for a greater mix of commercial and not-for-profit organizations engaged in electronic publishing of the results of scientific research.

A. University Support for STI Electronic Publishing

The key university strategy to contain the costs of managing STI is to introduce competition into the marketplace for STI by encouraging not-for-profit organizations to engage in electronic publishing of scientific research results. A more competitive marketplace requires that established and new not-for-profit organizations enter STI electronic publishing. These organizations could include research libraries, university presses, scientific societies and partnerships among these and similar organizations. Partnerships between universities and scientific societies should be explored as these institutions share many of the same objectives in knowledge creation, quality control, and dissemination at low cost.

By definition, not-for-profit organizations have an economic structure that makes them accountable to the organizations' membership or to the general public to operate on cost-based terms. This is an important characteristic when determining actual costs for a publication and for recognizing cross subsidies that may occur, for example, when a publisher starts a new journal. This is the kind of an assessment difficult to surmise from commercial publishing organizations.

The Task Force is aware of significant difficulties that not-for-profit publishers face when trying to compete with the for-profit sector. However, the Task Force sees an opportunity and a compelling self interest for universities to act to build and support a not-for-profit STI publishing base for the future -- a base essential to receive new STI intellectual property that might otherwise be assigned to commercial publishers perhaps because not-for-profit publishers did not exist.

Some not-for-profit professional societies (e.g. American Phytopathological Society, American Society for Microbiology) are moving towards electronic publication of print-based journals to contain costs. The combination of improved technology for mathematical (graphs, tables) and photographic (including electron micrographs) transmission, increased postage and increased paper costs are compelling. A transition period is likely in which both paper and electronic formats are available. (See also The Landscape of STI, in Appendix A, for examples of other professional society publishing programs.)

Electronic publishing of the results of scientific research is not an insignificant burden to ask universities or other not-for-profit organizations to assume. In some cases, the Task Force recognizes that additional financial support for such organizations is required to make this concept a reality. However, creation of competitive outlets for publishing STI remains the single most important action with potential for a long term solution to contain the costs associated with managing STI.

University leadership and financial support is required to build a cost-based niche in the STI marketplace to foster competition so that publishers of high-priced STI journals moderate their subscription prices.

B. University Actions re STI Intellectual Property

The current practice by scientists of assigning total intellectual property ownership to STI publishers is a barrier to achieving university cost containment for library materials budgets. To maintain and expand access to STI research resources while containing university costs, the task force sees advantages to develop university policies, practices, and mechanisms to retain in the not-for-profit sector some or all of the ownership rights of the intellectual property that is created by university faculty and staff. (For a description of possible ownership scenarios, see the Report of the AAU Task Force on Intellectual Property Rights in an Electronic Environment, especially part 5, Four scenarios for change, and part 6, Corollary issues for further deliberation.)

The National Science Foundation estimates that 36% of all scientific papers published in journals are written by U.S. authors (30% for physics journals). [8]There are no precise data to indicate what portion of those U.S. authors are university-based, but informed estimates suggest that about two thirds of U.S. scientific authors are based in academe. University actions to encourage or require these academy-based authors of scientific articles to retain intellectual property ownership in academe would channel intellectual property to the not-for-profit outlets that are needed to make the marketplace for STI more competitive.

A fundamental concept underlying the scientific communication process was lost during the commercialization of the scientific journal. That concept is the importance of the university community as a supplier of scientific knowledge. This knowledge has too great a value for achieving national and global educational and research agendas for its ownership and control to be relinquished to enterprises with different goals.

C. University Acceptance of STI Electronic Publications

University reward and recognition systems must embrace electronic publishing as a legitimate outlet for publication of research results. A strategy that promotes non-profit ownership and cost-based electronic distribution of scientific articles authored by the faculty of North American universities would enhance access to these writings in all non-profit settings. For this strategy to contain costs in a significant way, it must be developed as a viable alternative to publication in paper journals. It must also result in journals that can compete in terms of academic legitimacy and editorial quality with traditional paper journals.

At present, the promotion and tenure process for faculty in most research universities recognizes primarily the Classical model: untenured faculty need to publish their research in the most prestigious journals. Electronic journals need to be made as prestigious as classical ones if the Modernized and Emergent models are to succeed. This is a function of the editorial boards and editing process, not of the format of the journal.

To achieve the transformation of STI management, it is critical that university reward and recognition systems embrace electronic media as a legitimate outlet for publication of research results, and that institutions develop strategies to encourage electronic journals to achieve as high a level of editorial quality and prestige as is now assigned to paper journals.

Figure 3
Monograph & Serial Costs

V - A Role for the U.S. Government

The U.S. Government invests nearly $70 billion annually to support scientific and technical research. To use much of the results of this publicly funded research, scientists (or the libraries that support them) should not be required to pay prices set by commercial publishers at rates much higher than cost. [9]

To introduce more competition to the marketplace for STI and speed the dilution of the current monopoly-like nature of commercial journals as outlets for scientific and technical articles, the Task Force recommends that the U.S. Government assume a role in the transfer of publicly funded STI research results to not-for-profit channels. This would leverage the science agencies' investment of $70 billion by ensuring that subsequent research is less costly. It would make a significant contribution to a national strategy for managing STI, and will reduce the cost of access to STI for those government agencies that purchase scientific journals for their own use.

Agencies of the U.S. Government that fund scientific research should, as a condition of funding, require that a copy of any report, data sets, or articles developed as a result of publicly-funded work, be deposited into an electronic outlet that will provide public access. The concept does not prevent an author from simultaneously submitting the reports, articles, or data sets for traditional publication or electronic dissemination elsewhere. All that is required, assuming public funding, is to deposit a copy in a repository that offers public retrieval services at cost-based prices.

Proposed National Repositories

To carry out this proposal, the Task Force envisions a system of distributed national repositories to manage all of the functions traditionally performed by the paper journal for scientific articles written as a result of publicly funded research. The scope of these repositories might be defined by discipline or line of scientific inquiry, or perhaps by source of funding. The repository would be an electronic database that could be managed by a scientific society, library, university press, government agency, or any organization that shares the goals of the research and education community.

The research funding agency would designate one or more these national STI repositories as the point for registering a copy of the article, report, or data set.

Each repository would have two components: a Current Collection and one or more Permanent Collections. Both parts are network accessible electronic databases. The Current Collection contains reports in preprint form as in the Los Alamos electronic preprint distribution system for particle physics. (One difference is that the Los Alamos system is completely voluntary. Authors choose to submit their work to the system; there is no requirement.)

The Permanent Collection replicates functions now performed by the paper journal. Entry into a Permanent Collection requires peer review to validate the research. Any article or report adopted into a Permanent Collection will, as the name suggests, be archived permanently. To guard against the impermanence of organizations, contractual conditions that specify conditions for the transfer of a Permanent Collection may be necessary.

Submission of an article to a Current Collection repository does not affect an author's freedom to submit the same work to other electronic bulletin boards that may exist, and to submit it to another publication channel, such as a paper or electronic journal. If such a journal demands exclusive publication rights, then the author may request that the peer review process for the Permanent Collection be bypassed. In such a case, the work will reside in the Current Collection for two years and then be deleted, without ever being considered for inclusion in the Permanent Collection. But in no case (assuming some fraction of federal funding) can the author choose not to have the work be available in the Current Collection as a preprint.

Both Current and Permanent Collections will be made available at a reasonable cost to the scientific community for browsing, searching and document retrieval. Such availability may or may not be mediated by the library system in the user's institution.

As the nascent form of the majority of scientific documents is electronic, this proposal need not be a significant burden on authors. The goal is to allow the rapid dissemination of articles via cost-based channels, building national collections of digitized STI resources and lowering the costs for access and use.

The success of the proposal should be reflected in a reduction in the number of paper STI journals and in more competitive pricing for STI journals and article delivery services.

VI - Implications for Research Libraries and their Institutions

The Task Force findings have significant implications for the entire university, not just the research library. All participants in the system of scientific communication will be affected as the system is transformed. At issue is the customary manner by which scientific research and discovery are conducted and supported and through which the results are shared with current and future scientists and students.

The Project Steering Committee asked the AAU Task Forces for recommendations to assist universities to achieve a greater return from their current level of investments made in the collection and distribution of research collections. This Task Force concurs with the Steering Committee's interim assessment that the issues under scrutiny can no longer be addressed as "business as usual." Actions should be taken to encourage campus constituencies to recognize the benefit of change -- change in the library and elsewhere on campus -- in order that scholarly access to research resources may be provided at costs that are sustainable.

Short and long term strategies are necessary to effect change in long-standing university practices and commonly held assumptions and values. The recommendations presented in section I of this report are proposed next steps for AAU and its member universities. It should be understood however that these actions are intended only as the next steps in a strategic direction that will, in the long term, exert a positive influence over the transformation of scientific communication.

VII - Opportunities for the University Beyond the Sciences

There is extraordinary potential in the current electronic environment for innovation and discovery of new methods of scientific inquiry and for improvement in the management of STI. Indeed, the opportunities for the university as a whole surpass those for the sciences. Institutional and library actions that take full advantage of networks, networked information, and creative applications of information technology are strategic investments that

• Recognize the changing nature of scholarly research and communication and its reliance on networks and new technologies,

• Support research and teaching in all disciplines,

• Enable libraries to improve their services for STI and other research resources as well, and

• Provide universities with a capacity to pursue distance learning, independent education, worldwide education, and industry/education collaboration.

The challenge for universities and libraries -- individually and collectively -- is to achieve a level of readiness that will maximize the potential of such applications in support of science and to leverage this investment for the benefit of the institution as a whole.

Endnotes:

1. See Shirley K. Baker and Mary E. Jackson, Maximizing Access, Minimizing Cost: A First Step Toward the Information Access Future, prepared for the ARL Committee on Access to Information, (Washington, DC: Association of Research Libraries, February 1993); and Marilyn M. Roche, ARL/RLG Interlibrary Loan Cost Study, (Washington, DC: Association of Research Libraries, June 1993).

2. National Research Council. National Collaboratories: Applying Information Technology for Scientific Research. (Washington DC: National Academy Press, 1993), 7.

3. Efrat Livny, "Introduction: Human and Plant Genome Projects." Bulletin of the Medical Library Association 81 (July 1993), 273.

4. C. R.. Clauser et.al., "The Upper Atmospheric Research Collaboratory: A Testbed for Electronically Supported Collaboration." Transactions of the American Geophysical Union, in press.

5. National Research Council, 7.

6. Mary Biggs, "The Proposed National Periodicals Center, 1973-1980: study, dissension, and retreat," Resource Sharing and Information Networks, Spring-Summer 1984.

7. Report of the ARL Serials Prices Project (Washington, DC: Association of Research Libraries, May 1989).

8. International Science and Technology Update (NSF 91-309) (Washington DC: National Science Foundation, 1991).

9. A U. S. Government information policy that addresses the issue of cost for access to and use of publicly funded scientific research is "Data Management for Global Change Research Policy Statements," July 1991 [published in The U.S. Global Change Data and Information Management Plan, A Report by the Committee on Earth and Environmental Sciences, National Science Foundation, Washington, DC, 1992, pp. 43-48].