Access to and Services for Federal Information
in the Networked Environment

Access to and Services for Federal Information
in the Networked Environment

(Draft March 1997)

EXECUTIVE SUMMARY

Main Recommendations
This white paper recommends that decision-makers, including library directors, chief academic officers, and chief information officers, in organizations that rely upon and provide access to federal government information reassess their investments and program strategies to reflect the dynamic changes taking place in federal agency publishing and distribution. The old system is obsolete and there needs to be a re-engineering of federal information programs in light of network developments.

The paper makes three overall recommendations:

• Decision-makers need to reassess their institutional investments in and policies for selection, acquisition, access, service, and preservation of federal information in the networked environment.

• At the institutional level, collaboration is needed to bring together the range of skills necessary to provide networked federal information. At the national level, inter-institutional collaboration is needed to realize potential economies of scale.

• Given that access to federal information is a hallmark of our democratic society, institutions have a responsibility to advocate for federal information policies that will ensure continued access to networked federal information for all citizens.
  
  

Background
For the last ten years the federal government's focus on accountability, budget management, and the potential of rapidly developing information and communications systems has resulted in the development of policies and practices which are significantly changing how agencies create, produce, and disseminate their data, information, and knowledge. The pace of change has accelerated in the last five years and will continue to do so between now and the end of the century. Federal information distribution policy in the electronic environment is now more diffuse as agencies are becoming increasingly independent of the Government Printing Office (GPO). This shift is producing both opportunities and challenges for institutions who collect and service federal information.

The Problem
The problem is that what has been a stable, well-known system is now in flux and local institutional investments which have supported providing access to and use of federal information are increasingly out of sync with the future of federal information. The Net is in its infancy and is still evolving while, at the same time, institutions are grappling to provide a serviceable collection during this dynamic time of transition. To date, efforts to make networked information accessible often represent individual projects that may not reflect an institutional commitment to sustainability.

What This Paper Covers
Policy Directions
The evolution of federal policy regarding the distribution of federal information is now firmly on the path of electronic preparation and distribution. While there are continuing discussions about the pace of change and the continuing usefulness of print, the future of federal information production and distribution is clearly with the National Information Infrastructure (NII) and its assorted tools.

The important policy questions focus on how local institutions can adapt their own policies and strategic investments to capitalize on the opportunities created by this changing environment, as well as to establish discussions with federal agencies in order to build complementary programs.

Technical Directions
Federal agencies are adopting a number of technologies as they move their information to the electronic arena. Primary approaches include CD-ROMs and, increasingly, Internet accessibility particularly World Wide Web sites. These technologies have a wide variety of application possibilities, and hence, there is great diversity underlying what appears to be a consistent and coherent direction. In addition, the history of electronic federal information exhibits a variety of legacy application approaches that includes bulletin board systems, online manipulatible databases, flatfile databases, gopher sites, etc.

Technology questions center on how agencies can make their data available electronically so that users wishing to combine data from multiple agencies can do so seamlessly.

Production & Dissemination
Federal information production is increasingly electronically based, though federal information in print will continue to be a viable format for the foreseeable future. Not only is there wide variety among agencies in their application of information and communication technologies, there are also shifts in the value-adding processing (e.g., analysis and interpretation) which agencies have undertaken in the creation of their printed publications. Shifts are also occurring in the way federal agencies disseminate their information -- one of the more important is that many agencies are reviewing or applying fees for the acquisition or use of their publications.

The central production and dissemination questions focus on how local institutions will shift their investments in order to get federal information and make it useful in response to federal government policies -- as well as how to use local experience to inform federal decisions.

Use & Users
Today's information and communications technologies support new ways to mix, match, and manipulate digital multimedia information. User experiences and expectations are changing to reflect this pursuit of these new capabilities. Vast quantities of data and information are directly available via the Internet to a wide user community -- bypassing intermediaries and intermediary organizations. There are, however, an array of technical, policy, monetary, and human support challenges facing both the individual and the organization in the use of networked federal information.

The important use and user questions focus on the amount of new organizational and technical infrastructure required to facilitate access and use in a cost-effective way. Further issues highlight the need to collaborate with other organizations to jointly build a critical mass of organizational and technical infrastructure in order to spread the cost and the benefit across a number of organizations.

Implications
Collections
Traditionally when one speaks of federal information collections it is generally through three models of access: ownership, participation in depository arrangements, and partnership with federal agency programs. Networked federal information resources offer a fresh opportunity to rethink institutional collecting activities and to tailor collections to meet the needs of the user community.

• Institutions need to rethink what it means to collect federal information in the networked environment, leverage institutional strengths and resources through partnerships and consortia, and develop new models for collections.

Preservation
Preservation of information is the fundamental component of the archival function of a knowledge repository. It will continue to be a requirement in the electronic environment in order to satisfy user needs. Users will continue to expect that federal information which was placed in the "care" of the institution to be available, and they will expect that the integrity of that information will be assured. Preservation of electronic federal information raises new practical issues for institutions primarily because the information now becomes separable from the medium on which it may temporarily reside.

• Institutions need to form consortia or other cooperative arrangements to share the responsibilities and costs of preserving networked federal information; these consortia need to negotiate with the federal government the terms on which they will provide this preservation function.

Networked Information Discovery & Retrieval (NIDR)
Network technology offers many opportunities and challenges regarding what information is available to users and how that information is located. The network expands access for users and it changes the tools and strategies employed in the search and discovery process. However, mechanisms for locating federal information on the Net are rudimentary and less adequate than systems for other media. The issue of federal Web site authenticity is also an important one that affects research and scholarship. To date, there is no authoritative single point of entry for all federal information, and this is an important contributing factor to the haphazard nature of Net access. Sophisticated NIDR systems will not develop overnight and institutions need to develop strategies at the top level to deal with the inadequacies of today's systems.

• Institutions need to develop tools and network strategies that will provide users with an organized entry point to federal information, while at the national level they need to advocate for an authoritative access point and the development of standards that will facilitate network-wide indexing and representation of federal information resources.

Services
Institutions of higher education, public libraries, and state libraries have played an important role in providing information services for scholars, students, and citizens using the vast amount of material produced by the federal government. These organizations provide the expert knowledge, searching skills, and awareness of local information and communication patterns and needs to complement and sustain the technical and organizational infrastructure investments. These capabilities are absolutely essential to bring meaning to the present flux and inconsistency that characterize today's federal information environment. Networked federal government information will transform existing models of service which have traditionally been building-based and dependent upon staffing reference desks that serve a relatively defined user community.

• Institutions need to rethink their service policies in the networked environment, define the communities for which they will provide service, and develop new service models that embrace and exploit these new technologies.

Infrastructure

• Institutions need to plan for and invest in an infrastructure (equipment, connectivity, training, support, and financial models) that will allow their clientele to take full advantage of federal information in the networked environment.

Opportunities & Challenges
The potential of the continuing development of the NII is the vision of a world in which people can easily discover, evaluate, select, retrieve, use, and combine information resources in the widest variety of formats. Our federal government's information -- often the heart's blood of research and development, teaching and learning, advocacy and local government decisionmaking -- is increasingly available to users as part of the Information Superhighway. However, our institutions are grappling with myriad technical, policy, and monetary challenges to realize the potential of networked federal information.

It is essential that organizations with a critical mass of infrastructure and capability -- and a reliance on federal information or a responsibility beyond their organization to citizens -- have a working awareness and understanding of the opportunities and challenges that networked federal information offers. Organizational leaders need to position their organizations not only to take advantage of these new opportunities, but even more importantly -- to actively participate in meeting these challenges while creating solutions essential to the successful use of federal information.

INTRODUCTION

Overview
Federal information plays an important role in the mission of our research and education institutions. Federal information can be a critical piece of a research initiative, as well as a key element of the teaching and learning process. It can provide the foundation for scholarly research and/or provide the springboard for this type of work - from NSF guidelines for obtaining grants, to U.S. census data, to inventories of hazardous chemicals, to pending regulations, to databases of scientific information, to health care information, to diplomatic post records. Examples, by academic discipline, of some commonly used resources include:

History - Foreign Relations of the U.S. (FRUS); Weekly Compilation of Presidential Documents; Congressional testimony and reports

Demography/Public Policy - Census data; data from the Current Population Survey; Vital Statistics of the U.S.; Geographic Information Systems (GIS); Federal Register; Code of Federal Regulations

Political Science - FRUS; World News Connection; Congressional Record; Statutes at Large; U.S. Code; Congressional testimony and bills

Business/Economics - Bureau of Labor Statistics data; Budget of the U.S.; National Trade Data Bank (NTDB); Securities and Exchange Commission data; Treasury Department statistics

Environmental sciences - Toxic Release Inventory (TRI); superfund information; EPA environmental impact statements; wetlands surveys

These resources and more provide the primary source material that researchers, students, and citizens have depended upon for decades. The ability to identify, locate, and use this information enhances research productivity and student learning.

The way government information dissemination programs have operated until now has been in a structured and organized environment. The Government Printing Office (GPO) has traditionally been the required primary printer for most agency publications. The agencies pay for the cost of printing out of their budgets and, increasingly, they are resisting this federal requirement to use GPO since some feel they can publish their materials faster and more cheaply using in-house printing technologies or by contracting out their printing. At the same time, GPO is not the sole disseminator of government information. The National Technical Information Service (NTIS) as well as some agencies such as the EPA and NASA operate well established dissemination programs.

Federal agencies have issued their information in multiple formats for a number of years. The formats issue is not a new challenge facing users or institutions providing access to federal information. Electronic federal information is merely another in a long list of formats.

Users of federal information have traditionally accessed materials in a variety of ways. Much federal information has been available either at no cost or for minimal cost through intermediaries such as libraries, directly from the agencies themselves, or from clearinghouses. Acting on the principle that government information is a public asset and the cornerstone of a democratic society and informed citizenry, federal policies have provided for guaranteed access to federal information.

The provisions of Title 44 of the U.S. Code establishes and describes the Federal Depository Library Program (FDLP) operated by GPO. Providing federal publications through depository libraries is one of the ways the government keeps its citizens informed of its workings in a timely manner. However, not all materials are printed by GPO. Publications not printed by GPO generally do not make it into the FDLP and must be accessed in some other way.

The FDLP principles of ensured access, service, use, and preservation of federal information, although still valid, were established in an age of print materials and do not reflect the information technologies of today. Much debate has surrounded the updating of Title 44 and many issues remain unresolved. Legislation to amend Title 44 to incorporate the current technological changes is underway in the 105th Congress, but it is not yet statute.

At the close of the twentieth century, with the increasing use and availability of networked information technologies, there has been a significant change in how federal agencies disseminate government information. This change is resulting in new distribution mechanisms, as well as new and changing user needs and expectations. As a result, institutional leaders (library directors, chief academic officers, and chief information officers) need to rethink their responsibilities and capacities of their organizations in this shifting environment to facilitate the flow of federal information to academic and citizen communities.

The Internet (and its related tools) is an enabling technology with the potential to provide seamless access to information for scholars, students, and citizens. The Internet can offer people broader access to information -- if they have the appropriate technologies and telecommunications capabilities. It also offers the prospect for the development of new types of services that do not rely solely on a building-based delivery system.

Digital federal information will not immediately displace print, so institutions will likely have to invest in the financial support of more than one delivery approaches for quite some time. However, the movement to networked federal information and communications is permanent. This will require increased local investments in technical capabilities, equipment, software, telecommunications, and human support and mediation.

Not only do organizations relying on federal information have to invest in a variety of infrastructures and organizational capabilities, they must ensure that the information is accessible and usable. This goal can only be met if organizations initiate discussions with federal agencies to educate them to the needs of local institutions.

Policy Directions
Federal policies affecting public availability of government information arise from a wide variety of laws and regulations. Title 44 of the U.S. Code, for example, provides for the GPO Sales Program and the FDLP. Both programs provide for the distribution of federal information from a majority of federal agencies. The Paperwork Reduction Act of 1980 and its reauthorization in 1995, together with OMB Circular A-130, provide a policy framework for the management of federal information resources. This legislation coincided with the growing viability of the networked delivery of government information and the increase in the cost of printing and distributing federal government information. During the FY 1996 Legislative Branch Appropriations process, the conference committee called for GPO's budget submission to be consistent with the strategic plan (included in the study mandated by Senate Report 104-114) to assure substantial progress toward maximum use of electronic information dissemination. This strategic plan delineates a system for a rapid transition to an electronic government documents publishing and distribution system.¹

The evolution of federal policy regarding the distribution of federal information is now firmly on the path of electronic preparation and distribution. While there are continuing discussions about the pace of change and the continuing usefulness of print, the future of federal information production and distribution is clearly with the NII and its assorted tools. All of those individuals and organizations relying on federal information must address the infrastructure, services, and programs which will make possible the effective use of electronic federal information.

Technical Directions
"The design of successful technical platforms demands a synthesis of technologies and practices from computer science, computer-communications networking, information science, librarianship, and information management."²

Federal agencies are adopting a number of technologies as they move their information to the electronic arena. Two primary approaches include CD-ROMs and Web sites. Both of these technologies have a wide variety of application possibilities, yet there is great diversity underlying what appears to be a consistent and coherent direction.

The history of electronic federal information exhibits a variety of legacy application approaches that include an agency's use of Bulletin Board Systems, online manipulatible databases, flatfile databases, relational databases, gopher sites, etc. Data access presents different problems from text access. Technical compatibility, proprietary formats, data description and access standards, the use of standard languages like SQL -- are other examples of the many diverse approaches which are part of today's practice in the creation and distribution of electronic information.³

Most of these issues carry into the CD-ROM and networked environments as well.

"The important technology ... questions focus on how agencies can make their data available electronically so that users wishing to combine data from multiple agencies can do so seamlessly."⁴

There is much work to be done by both federal agencies and institutions of higher education, as well as state and public libraries to ensure that this essential goal is met.

Production & Dissemination
The rapid development of federal information policy and technical infrastructure related to the NII is resulting in significant changes in how federal agencies disseminate government information. Federal agencies are now exerting greater autonomy in their use of information and communications technologies for their information production and distribution. The Internet and its related tools provide opportunities for agencies to deliver services, provide communication channels, as well as deliver publications. Production is increasingly electronically based, though federal information in print will also continue for the foreseeable future.

As they consider their continuing use of information and communications technologies in their missions, agencies, and the federal government in general, are expecting that the red-tape costs of federal information will be minimized, while the utility of government information will be maximized -- and agencies will be passing on and sharing costs in new ways. Agencies are increasingly considering or applying fees for the use or purchase of their data, information, and knowledge. No-fee access (which has been the predominant approach of the history of the FDLP) is seeing some shifts as federal agencies begin to recover their costs by charging fees.

Not only is there wide variety among agencies in their application of information and communication technologies, there are shifts in the value-adding process which agencies have traditionally undertaken in the creation of their printed publications. More information may be available, but users could encounter raw data without the analysis and interpretation that many agencies publications have provided. This fact has significant infrastructure implications for institutions that provide access to this type of information.

The important production and dissemination questions focus on how local institutions will shift their investments in order to access federal information and make it useful in response to federal government policies -- as well as how to use local experience to inform federal decisions.

Opportunities and Challenges
Electronic formats can facilitate the potential visibility and utility of federal information by making it available to a wide audience. Many agencies have already put up homepages on the Web introducing visitors to their various missions and services.

The Net encourages active learning by stimulating curiosity and creativity. Users can experience learning in an engaging environment that can break down barriers which, in turn, can lead to further study and other intellectual pursuits.

The Net is an enabling technology and has the potential to provide seamless access to federal information for the scholar and student. The network offers users broader access to information, especially to those with no nearby library. It also offers the prospect for the development of new types of services, e.g. interactive customer-oriented services linked to publications. By embracing these technological changes, institutions that provide access and services can have an impact on the network's future by adding value to this information.

The infrastructure and delivery of information and services is shifting and poses new dilemmas, questions, and opportunities for institutions. The ability of an organization to provide its users with the access and services to which they are accustomed will depend on how the institution answers such questions as: 1) How will users get access?; 2) How will institutions continue to carry out their missions as preservers and servicers of information in pursuit of intellectual productivity?

In order to inform institutions of the challenges and opportunities for reaching this vision, this paper will examine the following areas and their resultant implications for the organization: Collections, Preservation, Networked Information Discovery and Retrieval, Services, and Infrastructure.

IMPLICATIONS

Implications: Collections
Overview
Introduction
Federal Agencies now consider information resource management (IRM), as expressed through the legislative mandates of the 1995 Paperwork Reduction Act, to be a critical component in their efforts to economize, streamline, and improve their vast array of programs and services. As an indication of how far the reliance on electronic data interchange has evolved, nearly all federal agencies either manage a growing package of electronic products and services, perform electronic adjudication and regulation oversight (by issuing rules and guidelines online), or offer direct citizen access/services through the Internet.

As a result of this network activity, Federal agency information resources are becoming more important to the conduct of the Agency's daily business. Agency staff no longer consider their information resources to be merely the printed by-products of the Agency's administrative processes This is a rapidly evolving movement away from the limitations which have been part of traditional printing and publishing practices. Naturally enough, this opportunity is profoundly altering the bibliographic and distribution arrangements that have been shared among the federal government, the scholarly community, and the nation's libraries.
Traditionally, when one speaks of federal information sought by scholarly and other institutions, it is through three models of distribution and/or access:

• Ownership -- through library collections and information centers;
  
• Participation -- through several different national systems of "depository" agreements and arrangements;
  
• Partnership -- through several federal government research programs (e.g., theNational Institutes of Health, the Department of Energy's national laboratories)

Networked government information collections offer a fresh opportunity to rethink collecting activities and to tailor collections more precisely to the needs of the local community. There is no doubt that, for the foreseeable future, existing heavily print-based research collections will continue to require service and preservation . Yet, increasingly, collections and users will depend on the full exploration and utilization of the possibilities offered by networked collections.

During this period of intense and chaotic transition, the only constants related to federal information are change and inconsistency. There is no one method of interaction, however, and institutions must consider new strategies, new relationships, and new investments which will ensure effective access to federal information.

Issues
There are a number of concerns about federal information collections in the networked environment. They include:

• No commitment to provide continued access to information published either by the agencies or by the institution/user community
  
• An abundance of raw data with too little analysis

Government-published reports frequently include significant analysis of the data collected. Now that raw data can be made available easily over the network, there may be less incentive for the government to produce such extensive analysis. For example, there may be no geographic report produced on unemployment in the U.S.; rather the user is presented with a mass of raw data.

• In the network environment "one copy" is not enoughCurrently, many federal agencies sell access to their network files to recoup some of the cost of making the information available. These agencies work frequently with GPO to offer depository libraries no-fee access. However, that access is often limited to a single password or workstation. This limits access to one user at a time. As many institutions have found, use of information sources via network access frequently surpasses use of the same resource when it is held within that institution. As a result, single-user access is frequently insufficient to meet the demand. To provide supplemental access, institutions need to reallocate funds to purchase additional access directly from the producing agency. STAT-USA/Internet offered by the Bureau of Economic Analysis is one example, and the replacement for the Foreign Broadcast Information Service's (FBIS)Daily Reports series, World News Connection, is another.
  The key change in this new environment has been the shift from a static environment to a dynamic environment. Until now most users have accessed federal information by "coming to it." Now they have the opportunity to interact with it.

  However, because of the Net's high capacity for data transfer and speed, institutions will need to redefine relationships with their community of users, with peer institutions regionally and nationally, as well as with the federal agencies responsible for producing the information.
  The most important feature within this electronic context is no longer ownership but, rather, access. No single institution (or institutions) will house federal information as has been done up to this point. In the distributed networked environment, there is not the same need for a set number of copies as in the traditional environment. However if one copy is not enough, how many copies are sufficient? Collection policies and priorities will need to be rethought in light of the changes brought about by the network.

  Federal agencies may continue to "validate" their data through their legal authority, but they may no longer undertake the value-adding processes which in the past resulted in recognizable "containers." Increasingly, individuals have to "add the value" that "fabricates" the federal data and information into products and services designed around their individual needs. Not only must institutions invest in the technological infrastructure that is part and parcel of access to networked federal information, but other organizational value-adding processes will require further investment. Because of this, scholarly institutions must assume a much more active role (including financial investment) in the creation of products and services that are designed to facilitate this fabrication.

  Success in managing federal information collections made available over a network will depend on:

  • preserving and providing access to segments of the electronic government information stream OR a willingness to depend on other non-library intermediaries for long-term access;
    
  • the ability of institutions to move information off the Internet and onto a local network or onto some type of media-specific format (print, microform, or electronic);
    
  • the willingness of institutions to coordinate the sharing of collection responsibilities for various segments of government information;
    
  • the ability of institutions to reallocate resources for hardware and software with which to store and manipulate network information.

  What might a networked collection of government information look like?
  Imagine a collection of statistical data like the Consumer Price Index or Leading Economic Indicators arriving in libraries and being immediately integrated with historical data. Users would find that monthly, quarterly, and annual information could be retrieved, then downloaded into the user's preferred format with equal speed and consistency. Statistics indexed to a base year could be simply recalculated to other base years on demand and used to make projections into the future.

  Codification of laws and regulations would occur simultaneously on the day they were scheduled to go into effect. Hypertext connections would enable the reader to review all stages of the legislative and rule making processes from proposal to enactment. Public comment on proposals or regulations, once too voluminous to be included in the printed format in anything but the summary version, would now be available full-text as part of the networked record. Via networked collections, it will be possible to monitor Congressional voting as it occurs, or even to participate in Congressional hearings from one's home computer.

  Networked collections will allow users to extract just the piece of government information they need, or to manipulate, repackage and annotate government information, so it will best meet their needs. In a networked environment, users will be able to take on research and analysis that was once too complex or time-consuming even to consider. Government will be able (cost effectively) to provide raw data sets to users who had been limited to studying the government's evaluation of a problem rather than testing their own statistical hypotheses.

  It might be possible in the not too distant future that community users would find on the Net a core set of government information resources with the full text of current and historical information provided by the producing agency or a government intermediary. A core collection set could include the opinions of all federal courts, the Congressional Record, the Federal Register, Statutes at Large, the Code of Federal Regulations, and the U.S. Code. Recurring tabular data sources, such as Vital Statistics of the United States, and recurring statistics on health, crime, education, immigration, agriculture, and even the current census are also likely to become core titles/information resources. Publications that are brought together by government-sponsored clearinghouses like ERIC might also become core titles for a limited time after their initial release.

  Institutions would grow accustomed to thinking of the core collection as part of their own collections. For these materials an institution's "collection" and its "network connection" would become synonymous.

  Current Situation
  A Snapshot Study
  Purpose:
  The purpose of this snapshot study is to determine the extent to which institutions could rely on Net (versus traditional) access for an agency's information/publications today. How do some agencies' Web sites compare with what they distribute through GPO? What are the implications for collections now? In addition, this snapshot is intended to evaluate how well these agencies are taking advantage of Internet capabilities.

  Methodology:
  Using the Monthly Catalog through GPO's Web site, in conjunction with the List of Classes of United States Government Publications, the number of titles published through GPO for selected federal agencies were identified for the time period January 1995-May 1996. This list was utilized to determine the number of print publications that were made available on the Internet. In the process, any enhancements were also noted.

  Results:
  The results are as follows:
  [Insert Chart Here] -- Currently located at the end of this section

  
  

  Bureau of Labor Statistics

  The Bureau of Labor Statistics (BLS) (http://www.bls.gov/) offers a prime example of the possibilities involved in launching a collection on the Internet. While the site does not include much of the analytical information put out by the Bureau, such as articles from the Monthly Labor Review and Occupational Outlook Quarterly, the amount of statistical information represented is quite comprehensive. Of the approximately 228 paper and microfiche publications released by BLS from January 1995 to May 1996, almost all were statistical publications. Therefore, the material that is most often required by the academic community and the public is available on the Web site. This information has been greatly enhanced by its presence on the Internet, as data can be manipulated in ways that are not possible in the paper format. With various methods of searching and formats of retrieval, it is possible for an expert researcher or an average citizen to find the depth of information needed. In the paper environment, a researcher may have to go to several different publications, and wade through superfluous information before the requisite data is retrieved, while through the Web site it is possible to create tables that are limited to the specific data requested by the user. Despite its enhancements, however, the collection on the BLS Web site can not be considered a true replacement for what the Bureau offers in paper format. Much of the analytical text that accompanies BLS data is absent from their Web site and in some areas statistical data may be lacking. It is possible, however, to see from the site how a complete collection could be represented on the Internet and greatly enhanced in the process.

  
  

  Census Bureau

  The Census Bureau Web site (http://www.census.gov/) is a site both for the academic user and for the general public. A series entitled, Current Population Reports, published by the Census Bureau, was specifically chosen to compare this print collection against the Internet site. Out of a possible thirty-three documents, twenty-six print publications are available on the Web site and twenty of the twenty-six are exact matches to the print collection, plus they are provided in PDF format. The information in the other six documents is presented in some capacity.
  Sometimes, however, the number of tables present in the print publication is reduced by half on the Internet site. Also, the extensive analysis in the printed publication is now only a brief synopsis (e.g. The Black Population in the United States: March 1995).

  There are instances when statistical information is not present even though there is a link to the information (e.g. Population Projections of the U.S. by Age, Sex, Race, and Hispanic Origin: 1995 to 2050). The Census Bureau Web site does have enhancements which allow an Internet user to create their own data files using the 1990 Census or to create a thematic map. The site also provides population clocks for the United States and the world giving a continual population count, lists the most current economic indicators, and contains the publication, Statistical Briefs , in PDF format.

  The Census Bureau site offers a subject approach to indexing the information from their publications, as well as other alternatives for searching the online documents by keyword, place, map, or staff. However, it is recommended to search for statistical information using the subject index. Also, this site contains information such as press releases, radio broadcasts, available publications and how to order them, and other general information.

  It should also be noted that the Statistical Abstract of the United States is available on this Web site. At present, an Internet user can search within the text of Statistical Abstract, but there are no links to the resources providing the data for the charts. The Web site indicated that many of the statistical publications were being provided through a subscription service which, for a limited time, would be free.

  
  

  Office of the President

  The collection of presidential documents currently on the Net (http://www.whitehouse.gov/) is in some ways an enhanced version of the print collection. Press releases, briefings, and speeches are all included in a single location and are searchable. There are "Briefing Rooms" which include basic economic and social statistics for all major categories with accompanying charts displaying the data. There is no comparable print source released by the Office of the President.

  In some instances, the general information offered by a print document is greatly supplemented by the information on the Web site. For example, in 1995 the Office of the Vice President released a small pamphlet on the GLOBE program. On the Internet, there is a full Web page for this program with significant in-depth information.

  While much of the presidential material is present and enhanced on the Internet, there are some areas of information that are clearly lacking. For instance, there were approximately ten documents published by the Office of the President on disability in both text and Braille versions. None of these documents can be found on the Presidential home pages. A further drawback to the collection on the Internet is that the documents that are present are often difficult to locate. In the print collection, all documents are published under the Office of the President and readily identifiable as such.

  On the Internet, each agency or commission under the president has its own home page often with different methods of searching or retrieval, with no common organization or classification as is present in the print collection.

  
  

  Central Intelligence Agency

  The Central Intelligence Agency Web site (http://www.odci.gov/cia/) is a site both for the academic user and for the general public. It contains its primary publications in PDF format: The World Factbook 1995, The 1995 Factbook on Intelligence, and The Handbook of International Economic Statistics 1995. Unfortunately, the other eighty-three publications released in 1995 through May 1996, are not available on this Web site.

  The site enhances its print publications with a number of features. For example, the list of Chiefs of State and Cabinet Members of Foreign Governments is updated monthly and indexed by country. Also, the site includes audio and video clips and photos to discuss the history of the CIA, to tour the CIA Headquarters, and to view an Exhibit Center which includes images with text about such items as an enigma encoding machine.

  However, other CIA publications which are of great interest to researchers such as maps, are not presently made available on the Internet. The site does provide ordering information for the maps and other secondary publications. It should be acknowledged that the CIA Web site provides a comprehensive list of its publications back to 1980 and in some instances back to 1971.
  Another useful feature of the site is its internal search engine for publications and public affairs information. The search engine allows the user to control the search query by selecting features such as case sensitivity and it provides assistance by giving helpful tips on formulating searches. The Web site also includes a suggested reading list, links to other intelligence information, speeches/testimony/press releases/statements, and general CIA information.

  Of particular note, access to this site was interrupted for awhile when hackers "broke into" the site and mounted a bogus home page. Once again available, the initial home page screens have warnings about it being an official government site and that unauthorized use is prohibited.

  
  

  Federal Bureau of Investigation

  The Federal Bureau of Investigation Web site (http://www.fbi.gov/) is a site that is informative for the general public. The site contains current press releases, photographs, descriptions, and backgrounds of the ten most wanted criminals, general facts and employment information, field office locations, frequently asked questions concerning the FBI, information on investigations (e.g. Unabomber), information on congressional affairs involving the FBI (e.g. Russian organized crime), public affairs information and information about the FBI training academy.

  However, this site cannot be recommended for use by an academic researcher. Only one of its publications is available on the site, FBI Law Enforcement Bulletin, from November 1994 to present. The main publication of interest would be the Bureau's Uniform Crime Reports. At present this valuable publication of compiled crime statistics is not available on this site. What is available is a press release regarding Uniform Crime Reports and a very brief summary of the extensive report.

  Considering that approximately sixty reports were published by the Federal Bureau of Investigation in 1995 through May 1996, having only one actual document on their Web site greatly diminished its usefulness. The FBI home page could benefit from such enhancements as an internal search engine, and an increase in the number of publications made available.

  Conclusions:
  This snapshot study of federal agency Web sites illustrates the future possibilities of a federal information collection on the Net. However, the results clearly indicate that, at present, traditional collections remain essential because full Internet representation is not yet available for these publications. In short, the digital collection today is still a supplement to the traditional collection.
  
  Potential Models & Strategies
  Prospects for Shared Government Resources in the Electronic Environment
  Sharing collections and collecting responsibilities appears to offer great potential for ensuring and safeguarding long-term access to government information. Ironically, sharing responsibilities for managing printed government publications collections produced marginal results and in many cases failed. Only with much difficulty have users been able to know, at the title level, the government publication holdings of other institutions. The complexity of government published material has often required a look at these material before one could ascertain its usefulness. These factors frequently led to a preference for having the collection in-house, even when this duplicated holdings of nearby institutions.

  Metadata describing networked government information databases will now offer users information about these complex products often down to the level of individual data elements. With networked access to shared collections, the interface can be made transparent to users and as timely as in-house access. The challenge now is the need for organizations to make and to fund long-term commitments to government information collections.

  Potential Models for Networked Government Information
  ICPSR
  The Inter-University Consortium for Political and Social Research (ICPSR) has been a significant player in collecting, preserving, and providing access to computerized data files. Its collections are machine-readable data files, the largest files being those produced by the U.S. government (election data, census, voting records and health statistical surveys). This model contains both advantages and disadvantages for the management of networked federal information collections. They are:
  

  • Components of the ICPSR model for collecting and preserving information might be a scalable model for managing certain types of government information.
    
  • ICPSR is a fee-based, membership organization with password restrictions that could potentially leave a large user group with no collection access.
  Organizations Working with Agencies
  One of the best examples of libraries partnering with government agencies is the ongoing arrangement between Cornell University's Mann Library and the U.S. Department of Agriculture. The project is titled the USDA Economics and Statistics System, (http://usda.mannlib.cornell.edu/usda/usda.html) funded by a USDA-CSREES grant, the system includes reports and historical data sets covering both domestic and international agriculture. This model, too, offers possibilities and drawbacks for federal information collections.
  
  • Users able are able to manipulate the information to suit their needs since reports are usually text files and data sets are available to download into a variety of statistical software formats.
    
  • The arrangement includes only a small segment of USDA's publications/materials, and it is unknown whether or not this is a scalable model.
  State-wide or Regional Sharing Programs
  Maryland's State Library Resource Center, Maryland's alternative to a state library, is administered by the Enoch Pratt Free Library. Coordinated network collections, including a wide range of government information, are made available via the state-wide public information network known as SAILOR (http://www.sailor.lib.md.us/). Cooperative programs such as SAILOR might serve as a coordinated mechanism for long-term access to selected government information in electronic formats. Some of the highlights of the SAILOR program include:
  
  • Funding assistance and training have been incorporated into the State Library Resource Center program, making it particularly useful and affordable for local county library systems to offer access;
    
  • Topical Area Review groups (TARs), including one for government information, have been formed to help shape the content and connections SAILOR provides to other libraries and collections;
    
  • SAILOR developed local interfaces, tailored to a Maryland audience, that enhanced access;
    
  • SAILOR "collections" of government information sources are a combination ofconnections to other network sites, as well as to connections to files mounted and maintained by SAILOR staff.
    
  ACLIN, part of the Colorado Information Network, provides free access to public and commercial information resources for Colorado residents. It is a central source for state government information, as well as for selected federal information. Components of the program include:
  
  • A defined mission which incorporates information resources that support the education, health, business, and social service activities of residents;
    
  • Toll-free dial access to the network for all residents whether they search the network from home, business, library, or school;
    
  • Cooperation and funding from both the public and private sectors.
  The question to be considered in both of these examples is, whether there is a commitment to continued funding of these programs and what would happen if funding ended?

  Strategic Partnerships with Federal Agencies: Citizen Access and Scholarly Organization through Electronic Reading Rooms
  The established system of federal depository libraries and/or information centers as distribution and access models needs to be rethought in the networked environment. Instead of building collections based on these relationships, scholarly institutions (and public and state libraries) could actively seek to build alliances and infrastructures with their peer institutions. Such alliances would allow for the identification, organization, and support of a common electronic interface for federal networked information or "electronic reading rooms." Electronic reading rooms, as a concept, would require planning, developing, implementing, and managing electronic "spaces" on the Web so that researchers and citizens alike could have predictable and stable intermediary places on the network to seek and find needed federal information without having to visit every agency Web site.

  Although these alliances should not, and could not, replace the legal and managerial responsibilities of the federal authorities to properly archive and preserve their information, they could be an important transitional device. In other words, these alliances would provide unique opportunities for both federal authorities and institutions to explore common problems and challenges in maintaining an effective public information life cycle.

  For instance, a particular university (or consortium of universities) could enter into a contractual arrangement with a federal agency to house and service the "host" on the Net. This arrangement would include the maintenance of a home page(s), answering general questions from the public, and creating a companion home page that directs people to other useful sites on the Net dealing with the same subject.

  In the case of the U.S. State Department, for example, this could include links to scholarly and popular sources of information about foreign policy, travel, international legal, public health, and environmental concerns. The institution would be responsible for maintaining the companion home page, and assuring the "validity" of the links maintained there. The State Department and the University of Illinois at Chicago (UIC), in a collaborative effort, have established the Department of State Foreign Affairs Network (DOSFAN) (www.state.gov/) to provide user access to a wide range of current foreign policy information.

  Community Information Organizations
  A community information organization, for the purposes of this paper, is simply an active alliance of community interests (rather than the interests of the scholarly community) with three primary goals.
  

  • First, through the coordination of common resources shared by community organizations (cultural, educational, economic), government offices, and scholarly institutions, it fosters the development of an open community computer information network; or, assists in the enhancement of an existing one.
    
  • Second, the alliance works to strengthen the organizational links between the public electronic network and public non-computer community information resources through the creation of necessary referral techniques.
    
  • Third, the alliances actively seek out, organize, digitize, and make available within the community computer network government information vital to the community's social, political, and economic well-being.
  Indeed, the wealth of information produced by federal, state, and local governments is often the key component within any community information organization. Wide distribution of this crucial "public knowledge" can make a difference in the economic development of neighborhoods and communities, the education of children and young people, the health of individuals, as well as the support of citizen participation in their local governments.

  Much of this information, along with the professional knowledge of how to accumulate, disseminate, and provide access to it, is found in the many academic, state and other public libraries that serve as document depositories for agencies at all levels of government. In particular, academic, state, and public libraries can provide critical elements of organization, information technology and telecommunication networking capabilities, as well as other institutional resources, that facilitated the distribution of government information across village, township, city, county, or school districts.

  Specifically, these partnerships will seek out opportunities to develop and implement networks of neighborhood-level public information resources that deliver a "basic package" of citizenship information services. Institutions can lend their considerable technical resources and connections to the Internet to foster the development of these community information networks.

  Information professionals, experienced in federal and other government information, would work toward the assurance that this "basic package" would draw upon a healthy mix of computerized and traditional community information sources from all levels of government.
  Recommendations
  

  • Institutions need to rethink collection policies to address the issue of what it means to collect in the networked environment and the opportunities and challenges that this presents.

  • Institutions need to form partnerships and/or participate in consortia that leverage institutional strengths and resources to develop new models for collections.

  • Institutions should leverage the strengths of the network by developing collections tailored to local needs.

  • Institutions need to develop strategies that bridge/link traditional and networked government information collections.

  • Institutions need to reassess the way in which they allocate collection resources, which traditionally have focused on ownership of government information, to ensure that there is sufficient access to networked information to meet the needs of their users.

  • Institutions need to develop strategies and mechanisms that will ensure long term access to collections important to their clientele.

  • Institutions should monitor federal agency Web sites for completeness and determine local need for retention and overlap with "traditional" federal information collections.

  
  Implications: Preservation
  Overview
  Introduction
  Preservation of electronic information continues the institutional mission embedded in the familiar paradigm: acquire information, organize it, make it available and preserve it. Institutions (particularly those in the FDLP) have participated in this significant, distinctive and successful role for print and other artifactual materials since the commencement of the program. In terms of fundamental principle or goal, there is no new issue, for the preservation role continues.

  Issues
  Why preserve electronic information?

  Preservation of information is the fundamental component of the archival function of a document depository. It will continue to be a requirement in order to satisfy user needs in the electronic environment. As this publication's several sections make clear, user needs will continue in most respects to be what they long have been. Users will want information reliably locatable, so that when they go there (whether personally or on the Internet) they can expect to find what they are looking for. Users will want information easily accessible: the location tools must be clear and accurate, and the information must be promptly retrievable. In the electronic environment the need for access tools will be more evident, and users will expect appropriate and standard software to be readily available. Finally, whether they are conscious of this need or not, users will expect information to be available that was placed in the depository's care a long time before; and they will expect that the integrity of the information they get from the depository will be assured.

  There are two broad strategic lines to be considered when examining issues of preservation of electronic government information: the technological and the organizational.
  
  Technological issues of electronic preservation
  
  As a matter of implementation, the preservation of electronic federal documents raises issues virtually identical to those libraries face in preserving other forms of electronic information. Indeed, the initiative of the Federal government in moving the FDLP to electronic provision may not be the first case in which the government has been the first stimulus for other public and private sectors to take steps that would eventually have been necessary in any case. As a result, institutional leaders and document program managers should keep before them that, while their present responsibility may be assurance of long-term access to government information, they are often likely to be setting precedents for the preservation and integrity of electronic publications of all kinds. These precedents, as we shall see, may be both technological and organizational.

  Preservation of electronic information raises new practical issues for librarians and archivists that did not previously have to be faced, primarily because the information now becomes separable from the medium on which it may temporarily reside. A book, or a printed Federal report, is published as an artifact. Like journals, manuscripts, sound recordings, CD-ROMs and other information resources which are published as objects that are their medium, artifacts exist in space and require specific physical handling to use. With such materials, to preserve the artifact is to preserve the information contained in it.

  In contrast, networked electronic information is volatile in two important ways. First, at the current time it always resides on media which themselves are fragile and have no demonstrated long-term life, even when compared with the low grades of paper sometimes used in government publications and certainly not when compared with the 300-year life of currently available acid-free papers. Magnetic tapes are known to be fragile, magnetic disks have not been seriously tested and CD-ROMs and other optical and magnetic recording techniques have when tested fallen short of claims made by manufacturers of even a few dozen years of life.⁵

  Second, electronic information is easily transferred from one medium to another with no loss, a technique with which we are all so familiar already that it has become senseless to talk of an "original publication" as opposed to a copy, for one (true) copy is as good as another for any practical purpose. The resulting ease of copying, of modification, of format change and of use is the positive side of such volatility. The negative side is the ease with which information can without detection be accidentally or intentionally lost or changed.

  One of the many consequences of this volatility is that, unlike with books where the decision may be postponed for years, preservation of an electronic document must be considered from the moment of its publication or even before if users are to be assured of its longevity and integrity.

  A very great advantage of the volatility of electronic information, and the lack of an "original" as such, is that the concept disappears of a "rare" or "fragile" or "unique" document which requires special care or protection from users. If a single (true) copy exists, many may quickly be made and users may use any of them without risking harm to any of them, even the indistinguishable "original."

  Information needing preservation
  
  In the artifactual environment, information is by definition published in static form. Narrative texts, tables of statistics, photographs and moving images are issued as they were produced at a point in time. Changing information (economic series, for example) are issued as serial publications or at least in separate print reports. Much current Federal information being issued in electronic form is of the same kind, and is likely to continue to be so.

  In recent years tools have been developed to support such reports, such as electronic databases (economic series, census data) and dynamic information resources (weather data and foreign travel advisories). These have served as bases for artifactual publications whether in print (labor reports) or CD-ROM (census data). The database and data collection tools, including their maintenance and preservation, have until now been the internal responsibility of the collecting agency. In the networked electronic environment both the tools and the reports are capable of being published, and the issue will arise of what the preservation responsibilities for the tools are as well as of the reports; and of whose responsibility it will be.

  For depositories as for research, state, and public libraries, it becomes more and more a responsibility to acquire and preserve databases underlying research as well as the consequent publications themselves. Publication of a database offers end users much more than merely a report derived from the database, as the user can manipulate the data with his or her own ends in mind; it is clear that depositories need to consider preservation needs for this class of material as well as those more similar to print.

  Dynamic information resources present a challenge of a different order: if the information flow provided by the government is endless, or nearly so (e.g. spacecraft telemetry data), choices must be made whether to preserve any of it, if so what to save, whether to use sampling or snapshot techniques, and so forth.
  
  Three kinds of electronic preservation
  
  Preservation of electronic information needs to be looked at from at least three points of view: medium preservation, technology preservation and intellectual preservation.

  
  

  Medium Preservation

  At any given time stored information is located on some medium, whether paper, magnetic disk, microform, punched cards, magnetic tape or chiseled stones. The artifact or medium will eventually decay, some more quickly than others. Medium preservation is the concern for preserving the specific medium on which information is stored. In the artifactual environment we have learned a good deal about proper environmental controls for paper, binding techniques, and the like.

  In the electronic environment less emphasis is placed on actually preserving the medium, though tape storage vaults also benefit from environmental controls. Instead electronic information is most often preserved from medium decay simply by copying the information from the decaying medium to a newer one of the same kind. This technique is called "refreshing" the medium; we speak of refreshing a tape by copying its contents to another similar tape. (In the current climate of protection of intellectual property rights, copyright concerns must be noted even when the intent is simply to make a replacement preservation copy.) Refreshing of information on microcomputer and server discs is most often accomplished through backup/restore techniques when a failure occurs.

  
  

  Technology Preservation

  More problematic than medium decay are the rapid changes in the means of recording, in the storage formats and in the software that allows electronic information to be of use. This is "technology obsolescence", and it is essential to recognize that the greater attention must be directed to the obsolescence of technologies than simply of the media**. These obsolescence include the means of recording, the storage formats and the software that allows electronic information to be of use.

  Rather than simply refreshing, we also need to speak of "migration": moving information forward through technology stages as they become available and as the old technologies cease being supported by vendors and the user community. At the simplest level, for example, it makes little sense to preserve a 200-bpi tape by refreshing it to another tape of the same obsolete density when it can be copied to a contemporary cartridge format or perhaps to a magnetic disc.

  Migrating information forward through computing and software technologies presents the greater migration problem. Files currently exist of information created on earlier microcomputers using obsolete word-processors and operating systems (e.g. WordStar running under CP/M on an Osborne). For optimal preservation of the substantive information, the data should be migrated forward to a current word-processor running on current systems. Similar cases exist for earlier computer-assisted design systems or economic databases, on mainframes as well as on micros.

  It is still a matter for discussion which of several options should be followed in migrating information:

  • migrate information through successive technologies as they appear, or migrate older information to a current technology only on demand?

  • migrate information forward to a successor application program, or maintain the original application program and use software technology to make it usable in later technology (there are intellectual property considerations here)?
  
  Each of these approaches has its own proponents, its technological problems and its costs, and experimentation will be desirable.

  What is not likely to be acceptable is the reduction of complex formats to simpler formats in the vain hope that the information will thus have a longer effective life. Reducing databases or interactive documents to simple ASCII format forces a loss of functionality -- nothing has been created in simple ASCII for many years, for good reason -- and in many cases will make the information excessively difficult to manipulate or otherwise useless to the intended user community. ASCII is no longer a lingua franca; it is a crippler of information.

  Similarly, it is unwise to propose reduction of data formats to a few "standard" formats in the hope that this will further use over a period longer than a very few years. Given formats may be upwardly compatible for some time, but a step-function of increased system functionality often has the effect of making previous formats suddenly non-functional or obsolete. Imaging formats have been particularly volatile in this respect, but even the most common database and word-processing formats have antecedents less than 15 years old, and it would be a bold administrator indeed who suggested (say) WordPerfect v. 1.0 as the "standard" or "simple" basis for preserving word-processing formats. There are few if any data structures beyond the 8bit byte that one can have confidence will still be supported 20 years from now.⁶

  
  

  Intellectual Preservation

  A third preservation requirement addresses the integrity and authenticity of the information as originally recorded. Preservation of the media and of the software technologies will serve only part of the need if the information content has been corrupted from its original form, whether by accident or design. The need for intellectual preservation arises because the great asset of digital information is also its great liability: the ease with which an identical copy can be quickly and flawlessly made is paralleled by the ease with which a change may undetectably be made.

  Here are some of the questions that arise for a researcher using electronic information: How can I be sure that what I am viewing is what I want to see? How do I know that the document I have found is the same one that another has read and made reference to in her footnote? How can I be sure that the document I now read has not been changed since the last time I read it? Note that in these cases backup is not the issue; rather, it is how we know which version we have or do not have.

  Three kinds of possible changes:
  

  Accidental change:

  for example data loss during transfer, accidents during updating, or saving the wrong version.
  

  Intended change (well-meaning):

  • New versions or drafts (authorial texts, legislative bills);
    
  • Structural changes: updating a vendor register or a Departmental directory;
    
  • Interactive documents, e.g. hypertexts (or Lotus Notes files) with note-taking capabilities.
    

  Intended change (fraud):

  The example of one's own work to cover one's tracks or change evidence; or of anther's work. Possible examples: political papers, laboratory notebooks, historical rewriting, legal documents, contracts. Solutions involve hashing techniques and cryptography (even though the end result is not encrypted but in the clear).

  Whatever technique is used must provide generality, flexibility, ease of use, privacy protection where desired, openness of documents where desired, and low cost. Most important, it must also function over long periods of time on the human scale, that is, after individual human actors are dead. Institutions providing access to electronic information will have to realize the extent of their commitment to assuring the integrity of the information they make available. With print this has not been a concern. In electronic form it always will be.

  Organizational issues of electronic preservation

  Electronic preservation requires new forms of institutional commitment because the organizational and fiscal obligations must be long-term.** Printed materials can survive loss of care for many years; electronic information can not. The Task Force on Archiving of Digital Information recently noted the significance of organizational issues when it said "the key that unlocks the path to the economies of the digital environment is not technological, but organizational".

  Organizational Commitment

  The permanent assignment of staff responsibility for the provision and long term maintenance of electronic information will be required. There is no single artifactual parallel for this responsibility: circulation, stack maintenance, preservation and physical plant departments now share it for print. Nor are there present parallels in academic computing centers, where staffs typically focus on technological advance and availability, leaving data to the users.

  Fiscal Commitment

  The permanent electronic depository will require assured continuity in operational funding. Almost any other library activity can survive a funding hiatus of a year or more. Acquisitions, building maintenance, and preservation can be suspended, or an entire staff can be dispersed and a library shut down for several years, and the artifactual collections will more or less survive. Digital collections however require continual maintenance if they are to survive more than a very brief interruption of power, environmental control, backup, migration and related technical care. Long term funding will be required to assure long term care. Institutions will need to develop new fiscal tools and use familiar fiscal tools for new purposes.

  Institutional Commitment

  The most difficult requirement will be that of conscious, planned institutional commitment to preserve that part of human culture which is in electronic form. The understood and willingly undertaken responsibility to provide institutional continuity will be the best assurance of the institution's ability to carry out its archival role. The institutional commitment will have to be clearly and publicly made if researchers and others are to have confidence that a given institution is indeed likely to exist for the long term. It is likely that adherence to a standard public institutional agreement to provide long-term electronic information will be required in the future for their libraries to be credibly regarded as archival repositories.

  Current Situation

  The current situation in electronic preservation can be described as one of preparation rather than practice. A few institutions have claimed from time to time to be maintaining long-term electronic archives, but have provided no substantiating information about standards they are following or of the formality of their institutional commitment.

  There are however several significant research and development efforts taking place in planning for the preservation of digital information which will provide resources and techniques for depository libraries seeking to provide permanent access to electronic government information. They include:
  
  The National Digital Library Federation (NDLF)
  Formed in 1995, the NDLF describes itself as "fifteen of the nation's largest research libraries and archives [which] have agreed to cooperate on defining what must be done to bring together -- from across the nation and beyond -- digitized materials that will be made accessible to students, scholars, and citizens everywhere, and that document the building and dynamics of United States heritage and cultures."

  The formation of the group was one of the most significant among libraries in years. Among important areas for discussion and development the NDLF has targeted is archiving: "Three important issues for NDLF consideration that surfaced ... are those of migration, certification of archives, and the fail-safe and/or rescue function." Others include rights, naming conventions, economics, interoperability, and discovery and retrieval. Some of its activities are documented in the newsletter of the Commission on Preservation and Access and at the Web site (http://lcweb.loc.gov/loc/ndlf/).

  Preserving Digital Information: Report of the Task Force (CPA/RLG)
  This report is also a significant milestone in library preparation for archiving issues, and involved experts from a umber of research libraries and elsewhere. At the end of 1994 the Commission on Preservation and Access (CPA) and the Research Libraries Group (RLG) created a Task Force on Archiving of Digital Information charged with investigating and recommending means to ensure "continued access indefinitely into the future of records stored in digital electronic form." The 21-member Task Force, co-chaired by Donald Waters, Yale University, and John Garrett, CyberVillages Corporation, have completed their final report, which is widely available online (http://www.rlg.org/ArchTF/) and in print.
  Studies in Scarlet: The Research Libraries Group
  RLG determined in 1995 that beginning a practical effort in creating a digital collection would help answer a number of questions that would arise. Their practical approach is focused, at present, on "Studies in Scarlet: Marriage and Sexuality in the United States and the United Kingdom, 1815-1891." It involves the digitization of existing materials rather than the creation of new electronic information, but anticipates having to deal with the many issues of long-term preservation in digital form as it sets up an archiving server to provide the information (RLG's Arches project). The RLG development is intended as a model for other institutions to follow rather than as a locus itself for a digital library. The Project is described on the RLG Web site (http://www-rlg.stanford.edu/strat/projdcp.html)
  The CIC Virtual Electronic Library
  The Committee on Institutional Cooperation (the Big 10 universities and the University of Chicago) have in the past two years developed the beginnings of their Electronic Journal Collection and what they propose to call the Virtual Electronic Library. The thirteen institutions are able to bring considerable competent brain power to bear on the collective solution of electronic library and archiving problems. Among other activities, they have formed a CIC Task Force on Preservation and Digital Technology, and have discussed a proposition to form an Ad Hoc Task Force on Federal Information explicitly to deal with electronic information. See the CIC Center for Library Initiatives Web site (http://cedar.cic.net/cic/cli.html).
  Consortial Efforts
  The cooperative nature of each of these projects underlines what is likely to be true for future efforts in electronic preservation: it will be a consortial or cooperative activity, with no single institution, however well situated, capable (or likely willing) to preserve all electronic information on its own. This will be true for Federal information as well.
  Potential Models & Strategies
  Locus of responsibility for preservation
  Two models suggest themselves for locating the responsibility for long term preservation of electronic documents: a Federal, centralized responsibility or a distributed, regional, shared responsibility.
  Centralized Federal Responsibility Model
  This model (suggested, in fact, by the GPO Report) would theoretically incur a lower social cost, as a single agency would be responsible for the systems and personnel costs associated with assuring longevity of the document record. It would have the advantage of simplicity, in that one would know where to go for the permanently available assured copy of a document.

  There are significant liabilities in such a model. It assumes a higher level of willing coordination between government agencies than has been true in the past, even in the print environment; without essentially perfect coordination electronically produced documents would fall between government cracks. It also assumes the willingness and ability of a single agency to take on such a task while at the same time assuring immediate and permanent access to the information preserved; it is by no means clear that government authorities see the latter complex and expensive task as their responsibility, nor whether there is an agency ready to take on the former task at the level required.

  Finally, the issue of credibility of government must be faced. Of greater popular concern, though probably lesser in reality, is fear of government malfeasors modifying information for personal or political gain or for Big-Brother-like manipulation. Of greater likelihood, particularly evident in the current political climate of reducing support for government functions, is an unpredictable hiatus in government funding at some future date placing in question the survival of all centrally-preserved documents. Though the likelihood of either eventuality is low, it is probably prudent not to rely on a wholly centralized protection method at current levels of confidence in the center.

  Distributed, Regional, Shared Responsibility
  A more practical and expedient model calls on a variety of public and private agencies to share the responsibilities and costs of assuring the survival of government information. Though the total social cost is higher, due to necessary redundancy and interconnections, the costs can be shared and thus may be more socially acceptable. (It should go without saying that Federal agencies, such as NARA and the Library of Congress, would be welcome participants in a shared scenario.)

  Redundancy is in fact a desideratum. Disaster protection is one obvious reason: a fire in a single institution's computer room will simply be less disastrous than a fire in a Federal archive that is the lone central archive of the nation's records (this is no less true today with the printed materials distributed in the FDLP, and there is no reason it would not be true for electronic information). Redundancy provides further assurance for document credibility; though integrity techniques can be implemented as described above, comparison of authenticated copies of critical documents may provide further levels of assurance.

  Multiple locations will facilitate speedy access. The balance between storage capacities at server locations and network bandwidths between them will shift frequently in the coming decades as usage grows and technology changes. Geographically-distributed centers of document preservation and access are likely to prove helpful as users all over the country search for government information. How many locations will be enough to satisfy redundancy needs for a given set of electronic documents? Two is probably too few, and 54 is too many. Experimentation, driven by local concerns, will provide the answer.

  The best reason for shared responsibility however is the availability of talent and energy in the institutions which will share it. It is desirable for many information professionals to be active in developing preservation systems, for it is presently impossible to predict which set of techniques will be the best. Library and information systems development has thrived best in the past because of multiple, competing development paths, and this is likely to be true in the provision and preservation of electronic government information.

  Principles of redundancy
  
  What principles should be followed in establishing multiple preservation sites (and thus access points)? Several models immediately present themselves:
  
  Broad subject categories
  
  "Topical depositories" could be based on science, agriculture, business, consumer information, medicine, and the like (networked information reduces considerably the need to think in terms of Regional Depositories). Several institutions with a particular interest in a given subject area might decide to work together to provide geographically distributed and secure sources of authenticated information. For example: Agriculture, for which Cornell University, the University of California at Davis, and the University of Illinois might take on the development of support systems and the provision of multiple sites for government information on the topic. The advantage of subject collocation for such material would be partially balanced by the disadvantage of not always being sure in which subject area a document might fall (agriculture or business?; agriculture or medicine?).
  
  Issuing agency source of information
  
  Archiving institutions might divide up the preservation of government information by the issuing agency, with some taking up NIH materials, some the Department of Agriculture materials, some the materials from judicial agencies, and so forth. An advantage would be a clearly understood set of responsibilities based on the issuing agency name.
  
  Pragmatic issues of file sizes and local institutional commitments
  
  Some institutions, particularly smaller ones, may wish to make specific agreements with other consortial members to share responsibility based on locally-defined issues and capabilities.
  
  Perceived responsibilities and abilities of agencies
  
  Federal depositories presently exist in many libraries: state libraries, state universities, private universities, large municipal libraries and elsewhere. Their different missions will lead them to participate in different kinds of consortia, either with like institutions or with institutions pursuing similar information goals.

  It is probable, and probably a good thing, that consortia for preserving and providing electronic government information will be formed on all these models, and on others as well. What is essential is that all the present institutions with federal information experience find ways to participate in the electronic federal information environment, for in fact the provision of government information to their constituencies is part of the mission of each one. Our nation will become the stronger, and our American citizenry will become more empowered, to the extent our nation's libraries and other institutions take up the preservation challenge of electronic information -- and of federal information.
  
  Recommendations
  
  

  • Institutions need to form consortia or other cooperative arrangements to share the responsibilities and costs of preserving federal information.

  • Consortia need to develop new paradigms for electronic preservation that leverage the strengths of information technology.

  • Consortia need to negotiate with the federal government the terms on which they will provide the preservation function for federal information.
  
  

  
  

Implications: NIDR
Overview
Introduction
There are many success stories that informed users of the Internet can share about accessing federal information. A political science student familiar with the White House's home page visits it regularly to read the daily press briefing. A demographer working on a population project is quite comfortable with statistical information available through the Census Bureau's home page. For individuals such as these, the Internet is a timely tool with relevant information to meet their needs.

Novice users of government information, however, are often intimidated not only by the bureaucratic nature of the federal government , but also by the complexity of its information. A citizen community group looking for environmental information on a planned playground that may once have been a toxic waste dump might find their search too broad for the Net to handle efficiently. In searching the Internet using existing tools, these users are often presented with a significant amount of irrelevant information. They can easily get lost while navigating the Internet, leaving them feeling frustrated and incompetent.

This section discusses issues related to Networked Information Discovery and Retrieval (NIDR) -- the mechanisms by which users locate, select, and retrieve information resources. Network technology offers many opportunities and challenges regarding what information is available to users and how that information is located. The network expands access for users, whether they are on-site or remote, and it changes the tools and strategies that they use in the search and discovery process.

The description and classification of information resources for organization, retrieval, and use is a well-understood (although difficult) problem with a long history. The networked information environment introduces a number of new considerations into description, classification, organization, retrieval, and usage. Among the challenges today are:

• the development of taxonomies for networked information;
  
• classification schemes to describe the content of networked information resources;
  
• evaluative information beyond descriptive cataloging which has been part of the bibliographic apparatus of print collections;
  
• the ability to describe widely varying levels of aggregated information sources and the granularity of information components that can comprise them;
  
• approaches for gathering the descriptive data for networked information.
  

Print-based federal information has its own set of approaches to catalog, index, and retrieve government publications. Networked federal information organization and retrieval systems, however, are in great flux -- in the same way that federal information creation and distribution are equally dynamic.

Issues
The Challenge of Federal Information
Federal information presents an unique challenge in the area of search and discovery. Users need to be assured that the federal information they access is authoritative. Government information is often the legal and regulatory language that determines such issues as the distribution of federal dollars (based on federally gathered statistics) or the guidelines for federally subsidized programs. The insertion of a small word such as "not" could entirely change the legislative/regulatory language and intent of a legal document and could, therefore, have far reaching implications for research and scholarship. The user, therefore, must be assured that s/he is directed to reliable sources of federal information.

The issue of federal Web site authenticity is an important one that affects research and scholarship. There is, as yet, no reliable, authoritative single (or distributed) point of entry for all federal information, and this is an important factor that is contributing to the haphazard nature of Net access for both end-users and intermediaries. In addition, a proliferation of Web sites which serve only as pointers is resulting in duplication of institutional efforts and confusion for the user. Although GPO Access is the legislatively mandated, centralized point of entry for electronic federal information, it is not, to date, the sole entry point nor is it a comprehensive site for all federal information.

The Search and Discovery Process
Using a variety of access methods to gather information, whether in the traditional or networked environments, is part of the iterative process of doing research. The researcher may first cast the net wide to gather as much information as possible and then refine and shape the nature of the research as experience and knowledge are gained. However, the research process should not be hindered by the access methods themselves. There is a great need for the development of seamless interfaces so that users can productively spend their time with the content of the information rather than trying to successfully navigate myriad, dissimilar information spaces.
The Search Process
In the search process, the user is trying to locate a known item or object whether it is a particular government report, a specific piece of legislation, a new regulation, or the latest figures on unemployment. There are several searching advantages to having information available in electronic form:

• full text or string searching allows the user to locate words or phrases within the text of a work;

• algorithmic ranking of documents matches word or phrase frequency with the user's inquiry and then ranks the results;

• thesauri provide controlled vocabulary entry points for users and also have the capability to automatically modify queries during retrieval;

• tree searching or other links automatically and seamlessly parse users' inquiries. (Marchionini p. 36)

What happens when the user is looking for a specific statistic such as the per capita income of the U.S.? In the traditional environment, the user could easily find this information in the Statistical Abstract of the United States ; provided that the user knew of this resource, was able to locate it through a tool such as American Statistics Index (ASI), or consulted an information specialist for assistance. On the Internet, this can be a frustrating and lengthy search that can be further compounded if the user does not know which agency produces this statistic. Depending on the search engine is used to execute this keyword search (e.g. Infoseek, Lycos, Yahoo, etc.), the number of hits could range from 29 items to 40,000 items. The user must then wade through the results hoping to find the statistic. To complicate matters, there are many false drops and while numerous references define per capita income, the statistic itself was either buried in the retrieved list (not in the top fifty) or not retrieved at all. A statistic that took only a few minutes to locate in the traditional environment, can now stretch into fruitless hours of searching via the Internet.
Discovery
Browsing is an integral part of the research process. Serendipity, exploration, and contextualization support the connections between known concepts and their linkage with new ideas. Who hasn't browsed bookshelves and happened upon relevant information to satisfy curiosity and/or an information query? Take the example of a researcher beginning a search on Medicare reform. A traditional subject search via the Monthly Catalog (MoCat) might be the place to begin, leading him/her to such agencies as Health and Human Services, Congress, the Executive Office of the President, and even the Census Bureau. The researcher might also want journal articles and other treatises and s/he would, therefore, also search a library catalog, a specialized index, and/or a commercial service like Lexis/Nexis. Footnotes and bibliographies would lead the researcher to still more resources as the search is shaped and refined through experience.

Using a broad search engine such as Yahoo or Alta Vista to begin a search, might be analogous to using a general index such as the Reader's Guide . The user might find some government information resources using each search engine, but not the wealth of subject specific federal agency materials that resources such as MoCat or CIS's Congressional Masterfile index. These indexes provide access to authoritative federal information that, for example, a teacher can browse for use in classroom instruction or can put a citizen in touch with federal information resources to meet a specific need. The networked environment offers the possibility of enhanced browsing provided that the user first finds an appropriate set of federal information materials. Then s/he can use hypertext links to find additional relevant material.
Current Situation
Today users are getting access to the federal information they need in a variety of ways using a combination of Internet and non-Internet methods.

• Direct access can be obtained through:
  
  • the Internet via a search engine;
    
  • knowing the URL of an agency or of a specific item through some mechanism as word of mouth, via a newsletter or listserv, etc.
    
  • direct agency contact, or through clearinghouses, such as the National Technical Information Service (NTIS) or the GPO Sales Program
  
  
• Many users, however, are going through an intermediary:
  
  • a federal depository or library, relying on the expertise of the information specialist to assist them in searching either/both the Internet or other traditional and electronic non-Net indexes.

While the direct access method works well for known item searching, the intermediary route provides the user with additional strategies for subject-oriented searches. Users should be able to find the information they need without knowing a URL to locate it.
What does NIDR mean in Federal Information?
In the traditional, institutional setting the search and discovery process for federal information has been enhanced by a variety of tools beyond MoCat (e.g. finding aids, catalogs, guides, and indexes) that organize and locate the information for the user. These various types of tools can be used at different points in the search process:

• for a known item search the user might consult MoCat and/or a source such as Commerce Clearinghouse's Congressional Index ;

• for a broad subject search the user might first consult a general source such as Congressional Quarterly's CQ Almanac and then move to a catalog and/or index;

• for a narrow subject search the user might directly consult a specialized index such as CIS's American Statistics Index (ASI).

It is perhaps because of these time-refined and well-controlled information resources and services that users have grown accustomed to using that have contributed to the high, if not higher, expectations of easy information retrieval via the Internet.

For federal information to have any value to the user, it needs to be organized and retrievable. The key components of any information search are:

• whether the information retrieved is relevant;

• whether all relevant material has been retrieved.

It is these two points, recall and precision, that express information search and retrieval performance. Traditional search methods to find federal information often takes patience and persistence as the user consults a variety of tools, but Internet technology, to date, only exacerbates the problems of search and discovery for federal information. Several reasons for this are:

• Networked information resources are extremely heterogeneous in nature, volatility, and coverage. They include a wide range of services and types of objects. This is part of what makes the NIDR challenge so difficult for federal information.
  
• Users need to be able to view the available information through a seamless interface rather than as a large number of collections organized by types of resource (e.g. gopher spaces, ftp sites, Web sites, etc.) or by methods of access.

What is needed to improve searching in the networked environment?
What are some of the issues that affect searching on the Internet? Referring back to the per capita income scenario, if the user knows which agency produces per capita income s/he could go to that agency home page and find the statistic. Since the statistic is available on the Internet, why wasn't it found when doing a keyword search? Some of the contributing reasons are:

• First, it depends upon how much of the Web site is indexed - full-text (few do this) versus URL, title, etc.

• Second, does the search engine used index only Web sites or does it also search gopher spaces and ftp sites? The Bureau of Economic Analysis, which produces per capital income, has, at this time, a gopher site in addition to its Web site.

• Third, there has been little progress in the indexing of non-textual materials. "Most non-textual objects are located through textual descriptions or linear scanning." (Marchionini p. 36)

• Fourth, what is the search engine's capability to rank items using statistical/probabilistic algorithms and then to return relevant results? These algorithms use word and/or phrase frequency to match queries with items. Additionally, what does the user need to know about this concept to search efficiently (e.g. Clinton and Bosnia Bosnia will return more relevant results in some search engines than just Clinton and Bosnia)?

• Fifth, what are the options within the search engine for the use of controlled vocabulary (i.e. thesauri) to find variant forms of words or to contextualize concepts (e.g. when searching for death penalty, documents with capital punishment should also be returned; per capita income might also be listed as per capita personal income)?

• Sixth, if the user can get to the appropriate agency site, does it have internal, searchable indexing? The Census Bureau's site offers searching by subject, keyword, and geographic location; not all federal Web sites offer this enhanced feature.

Search engines rely on information about objects. As more sophisticated NIDR systems develop, there will be increased emphasis on metadata. The good news for federal information is that the Government Information Locator Service (GILS) is setting criteria for agencies to develop information about objects or metadata and this standard will provide the underpinnings for improved search and retrieval.

Potential Models & Strategies
Where does this leave us?
Who is planning for network-wide federal information indexing? There is no current service that has organized appropriate and significant Internet and Web federal information resources that also offers the value-added structure, context, and level of specificity and description with a browsable scheme. GPO is attempting to do this with its evolving Pathway Services project. Its Browse function offers three features - Browse Titles, Browse Topics, and Browse GILS records; the Search function offers two features - search MoCat , with some URL hot links to federal sites, and the Pathway Indexer, utilizes a robot that crawls federal Internet sites building a database that allows keyword searching.

CIS and other commercial publishers are also entering this arena. CIS and Lexis/Nexis have recently developed a product that will provide access to digital government information. The Compass Library of Government Information will initially focus on legislative materials and will expand to include statistical, regulatory, judicial, and executive branch materials. Users will gain access through CIS's Web site which will apply a common user interface with content-specific help. Users will also be able to manipulate and combine data for incorporation into their own work.

Much traditional government document searching has depended on knowledge of agency structure and publishing, but information professionals should be able to leverage the strengths of technology in the development of efficient electronic search and discovery tools. At this time however, mechanisms for locating federal information on the Internet are rudimentary at best and are less adequate than systems for other media. Organization and indexing are chaotic making access haphazard for the user. The user must first be able to identify if there is networked information available to satisfy his/her need. Once that piece of the puzzle has been determined, the information needs to be in a usable format.

Sophisticated, thoughtfully conceived NIDR systems will not develop overnight. Institutions need to develop strategies at the top level to deal with the inadequacies of today's NIDR systems.

• There needs to be a designated, authoritative first entry point. The user could then be assured that authentic federal information would be found on that site.

• The site should also have well developed tools to use for finding information. First, there needs to be good menus to guide users in the initial phase of the search process. Second, there should be powerful searching mechanisms that permits parsing by subject and agency for efficient searching.

• There is also a need to go beyond doing Boolean searching on the URL, the title, and the summary text. If, for example, a user is looking for per annum wheat production and that statistic is not listed in one of those fields, the information will not be found.

• The user should also be able to refine searches performed by further using the results already obtained. This could be accomplished through Boolean, proximity, and fielded searching capabilities. Total natural language searching or artificial intelligence (AI) will not lead unsophisticated users to what they need.

Institutional strategies to assist users also need to be developed. There is a need for information specialists to create customized indexes and guides by subject, geographic location, etc.

• One strategy would be for institutions to develop one well-constructed home page that could serve as a gateway to authoritative, top node, federal agency sites.

• Another strategy, would be for an institution (or group of institutions) to identify the major subject interests of its user community and then develop front ends that might mirror, point to, etc. specific federal information sites that could fulfill the majority of the information needs of their clientele. For example, a citizen community networking coalition might provide access to such sites as the home pages of the White House, the Social Security Administration, the EPA, the National Center for Education Sta