23

contractual arrangements, it might be possible to integrate this data effectively with the data in the legislative information system through the design of the interface linked with the necessary security systems, and still remain within the requirements of the contractual obligations. There are a variety of options for accomplishing this, depending upon the flexibility of the vendor systems and the interests of the commercial providers. (See previous discussion under DATA SOURCES: COMMERCIAL AND NON-GOVERNMENT)

The power and flexibility of interface software can pose certain challenges. There are a variety of commercial providers of this type of software, and it can be customized by the user as well as by the designers of the legislative information system. The Working Group would need to establish some guidelines regarding the compatibility requirements of the software that offices could select to access the system. Absent such guidelines, the costs to develop and support all the possible software choices could be quite expensive.

The Library recommends that the Working Group form a subteam to design user interfaces(s), taking into account the needs of different users and the variety of data formats. The projected rate of technical enhancements in user interface software, the changing requirements of users, and the growth in the amount and kinds of data available on the system will require that the subteam make adjustments and improvements in the user interface(s) on an ongoing basis. The user interface should encourage and enable user to send suggestions for system improvements to the Working Group.

USER TRAINING, DOCUMENTATION, AND SUPPORT

Adequate user support through training, documentation, and troubleshooting is an essential requirement. Like most organizations in this country, Congress is a beneficiary of the dramatic improvements in the sophistication and in the ease of use computer technology that have occurred in the last five years. However, as this technology has become more widely used and as it is applied to larger number of critical business operations, maintaining ease of use has become more challenging.

The Library recommends that the design and support of the legislative information system follow these principles:

1. Design the system so that required training time can be kept to an absolute minimum.

2. Provide multiple ways for users to obtain the required training, including training classes, online help, printed documentation, self teaching programs, and audio and video instruction programs.

3. Through the Working Group, coordinate the work needed to provide user support, based upon the current programs and expertise of each of the organizations within the Legislative Branch. For example, the House Information Resources, the Senate Computer Center, GPO, and the Library each

24

provide training programs for their congressional clients. It might prove most cost-effective to divide and coordinate responsibility for introductory and advanced courses and for documentation among these organizations.

PRODUCTIVITY TOOLS

An important component in the development and improvement of a legislative information system should be an ongoing, coordinated program of evaluation, testing, and implementation of tools that can improve productivity both for creators and for users of the system. For example, under the distributed system proposed in this plan, various groups would be responsible for the preparation and tagging of data to standard specifications. A new piece of software that can make this process more efficient will benefit everyone. Similarly it would be valuable to encourage the development of automatic retrieval features that would benefit all users, such as allowing them to specify whether they wanted to search for an exact phrase, to search comprehensively for all variants of the words in a phrase (e.g., balanced budget, balancing the budgets, bringing the budget in balance with..., etc.), or search effectively for variations in the expression of an idea (e.g., ...ensure that federal expenditures match federal revenue generation...).

The selection of productivity tools for evaluation should be the result of a thorough understanding of the requirements of data creators and of users, balanced against technical opportunities. The basic principle is that resources spent on this effort should be business driven rather than technology driven. However, the testing program should be carried out by staff who have both a Hill-wide view and who also can gauge the potential of emerging technology to meet a variety of Hill-wide requirements.

RETIREMENT OF LEGACY SYSTEMS

The retirement of the systems that currently provide Congress with legislative information presents all of the challenges that arise whenever there is a proposal to replace an older system (or so called "legacy system") with a newer system. Despite their shortcomings, the Library's SCORPIO system, the House MIN and ISIS systems, and the Senate LEGIS system work every day for hundreds of staff on Capitol Hill and in district and state offices.

The basic principle should be that the current systems can be retired when the new system provides comparable functionality. The difficulty in applying this principle arises when one tries to specify exactly when comparable functionality has been achieved. Despite the fact that new technology provides many advantages over the older systems (e.g., the graphical user interfaces make them easier to understand, the user can "click" and link immediately to related information, such as the text of the debate in the Record cited in the status section of the bill), the older systems have been developed and fine tuned for over 20 years, and they have a number of features that are not yet available in the newer technology.

Here are two generic examples:

25

Systems operating on the World Wide Web cannot yet combine the results of previous searches with the results of later searches. There are ways to get around this problem, but the optimal and comparable solution is not yet clear.

Many expert users prefer to type in the specific system command rather than go through a series of "dialog boxes" because it is faster and more efficient for them. Most new systems permit this in limited and/or complex ways.

Here are two examples specific to individual systems:

The Library of Congress SCORPIO system allows congressional staff to request books located through its online card catalog through the ORDER command.

The House MIN system allows users to search for and read a summary of floor proceedings within minutes of their occurrence.

In the case of both of these specific examples, it will be possible to develop these capabilities in the new system, but it will take time and resources to accomplish the task. The new technology does not come already equipped with these customized features that have been developed over the years.

The older data accessible through these systems present an additional challenge. The data from the 104th Congress will certainly be accessible. However, the digests of the bills back through the 93rd Congress may not be transferrable to the new technology, although it will probably be possible to provide some limited means of access. While the older data is not searched as frequently as that of the current Congress, its availability, however infrequently needed, is still important to many staff. As one example, committees often ask CRS for information about historical trends in legislative activities. These questions can only be answered at a reasonable cost by accessing the older data in the current House, Senate, or Library legislative systems.

It is important to understand the significance of this problem without overdramatizing it. While nearly all of the data in the current systems exists in some print version, much of the value and accessibility of the data exists only through its availability in an online retrieval system. In this context, the definition of comparable functionality becomes a matter of judgment tempered by the issue of cost. At this point, it will require more analysis in order to make a recommendation about how to handle older data. Solution(s) to this problem will affect the decision about when comparable functionality has been achieved.

26

PRESERVATION AND LONG TERM ACCESS

The issues of preservation and long term access will be important for the new legislative information system. Much of the data in that system may exist only in digital format, and the question of how to ensure the preservation and long term access to the data in the new legislative information system must be answered before it is necessary to move on to the generation of technology beyond today's. Except for the problem of acid based paper, the invention of the printed book provided a solution to this problem for over 500 years. The current online systems have been in operation for less than 25 years and we now face the problem. Fortunately the advances in technology and the development of data standards such as SGML offer potential solutions to this problem in the future, but only if they are purposefully and effectively implemented from the beginning.

Because of the important role legislative histories play in the interpretation of Federal laws and regulations, the Working Group's design for a new legislative information system must plan for both the permanent archiving of core legislative data and its long-term accessibility. Legislative history, at whatever level of detail recorded, constitutes the framework from within which future legislative, judicial and executive actions will be taken, whether these are in the form of new legislation and amendments, court decisions, or agency rulemaking. For this reason, practically all legislative data collected must be indefinitely retained and constantly available.

An indefinite period of retention has the added benefit of meeting scholarly and public interest in the legislative process. Historical data can migrate to appropriate agencies with archival responsibility at a designated time and still have the potential to be integrated at the user interface level within the distributed computing environment that constitutes the legislative information system infrastructure. For purposes of security and reliability, certain data sets, particularly those in heavy demand, may be 'mirrored' or copied to more than one site.

OPTIONS FOR PUBLIC ACCESS

PL104-53 directed the Library to examine issues regarding efficient ways to make legislative information available to the public and to submit its analysis to the committees for their consideration and possible action. The Library interprets this issue to center on the question of access to the public without charge. There are, as noted above, a number of commercial providers who make legislative information available to the public for a fee. Also as noted above, the proposals in this plan, if implemented, would improve the quality, accuracy, and timeliness of core congressional legislative information available to the public through these commercial sources. The GPO ACCESS system and the LOC THOMAS system, on the other hand, are examples of congressional mandates to make legislative information accessible to the public without direct charge. The remainder of this discussion will focus on the issue of free access to the public.

27

There are three fundamental questions inherent in the issue of public access to Congress' legislative information system: 1) Should the public have access to the same public legislative data? 2) Should the public have access to the same technical system? and 3) Should the public have access to this information through more than one congressionally supported agency or office?

1. Access to the same core legislative data?

The answer must be yes. A fundamental tenet of democracy is the importance of an informed electorate, and the Congress has passed a number of laws to ensure that the public has access to government information. With the recent advances of technology that affect how government information is prepared and distributed, however, it is important to note the implications of this principle for the proposed new legislative information system.

A. Implicit in this principle is the need to ensure that the legislative information available to the public is as accurate and as timely as the information available to the Congress through this system. Technically this goal is achievable. The obvious qualification inherent in this statement is that if the information is not in the system it is not accessible either to the Congress or to the public. Committee chairs, for example, can choose not to make some information regarding committee actions available to the system. Despite the improved access that the public now has to legislative information through GPO ACCESS and LOC THOMAS, the most persistent criticism by some members of the public has concerned what is NOT YET available in these systems rather than what is. The implementation of the recommendations in this plan should address many of these concerns, but it may never be possible to address all of them.

B. Restricted and licensed information would be excluded from public access unless the contract specified otherwise. As noted earlier, many Members and staff find it valuable to subscribe to commercial sources of information in order to carry out their work. This plan suggests a number of ways in which that licensed information could be linked to the proposed congressional legislative information system so that it could be more easily accessible to offices which subscribed to it. Such information could be made available to members of the public who chose to purchase it as well, but it is not reasonable to assume that it could be made available for free.

Similarly, restricted material, such as CRS Reports, which CRS is prohibited from disseminating to the public, or other internal, confidential materials would be excluded from public access.

2. Access to the same technical system?

If the answer to question 1 above is yes (i.e., the public should have access to the same core legislative data on a basis that ensures the same accuracy and timeliness), then the answer to this question is also yes. The simplest way to