Networked Resources and Services: Next Steps


Copyright 1991 CAUSE From _CAUSE/EFFECT_ Volume 14, Number 2, Summer 
1991. Permission to copy or disseminate all or part of this material is 
granted provided that the copies are not made or distributed for 
commercial advantage, the CAUSE copyright and its dateappear, and notice 
is given that copying is by permission of CAUSE, the association for 
managing and using information resources in higher education. To 
disseminate otherwise, or to republish, requires written permission. For 
further information, contact CAUSE, 4840 Pearl East Circle, Suite 302E, 
Boulder, CO 80301, 303-449-4430, e-mail info@CAUSE.colorado.edu


          NETWORKED INFORMATION RESOURCES AND SERVICES: NEXT STEPS
                            by Paul Evan Peters

************************************************************************

Paul Evan Peters is Director of the Coalition for Networked Information,
an activity of the Association of Research Libraries, CAUSE, and EDUCOM.
He is currently vice president/president-elect of the Library and
Information Technology Association (LITA), chair of the National
Information Standards Organization, principal representative of the
American Library Association to Accredited Standards Committee X3
(Information Processing Systems), and a member of the editorial boards
of Library Hi-Tech and Research and Education Networking.

************************************************************************

ABSTRACT: This article examines the origin of advanced networks and 
explores why there is so much excitement about them among information 
technologists and librarians; explains what is meant by networked 
information resources and services, how they will ameliorate pressing 
library concerns, and what must be done to benefit from them; and poses 
four questions for the future that are fundamental to realizing the full
promise of networked information resources and services.


The often-predicted and long-awaited transition from information
distribution and access by exclusively print means to information
distribution and access by both print and electronic means now depends
upon a variety of institutional, organizational, and marketplace
"readiness factors" more than it does upon any specific technological
innovation and development. Accordingly, information technologists and
librarians will need to focus their combined talents and energies on
these readiness factors, so that this transition can be made as soon as
possible in the 1990s, and set the stage for the emergence of a system
of distributed digital libraries as early as possible in the 21st
century.

It is important to place the contemporary scene in the context
provided by the fifty-year (more or less) effort to marshall information
technology to the service of scholarship and pedagogy. Doing so helps us
to keep in mind what this long-term effort is really about. It is not
about "electronic libraries," "virtual libraries," or even "distributed
digital libraries." These popular and evocative phrases say something
about the technological and service architectures that shape the efforts
and aspirations of contemporary information technologists and
librarians, but they say nothing about what really motivates those
efforts. The mission of all of these efforts, no matter how
technologically or bibliographically esoteric they may appear to be, is
to improve information distribution and access by using high performance
computers and advanced networks to support research and education
communication.

Why are so many information technologists and librarians so excited
about advanced networks in general, and about BITNET, the National
Science Foundation Network (NSFNET), the global Internet, and the
proposed National Research and Education Network (NREN) in particular?

There are three basic reasons for this excitement: simplification,
connectivity, and performance.

First of all, an advanced network provides a common framework by
which to interconnect and to inter-operate the great variety of highly
heterogeneous departmental, institutional, regional, and other
individual networks that have sprung up by one means or another over the
last twenty years. This interconnection and inter-operation results in a
major technological simplification of the global networking scene and in
the reduced costs and the increased values that always accompany such
simplifications.

The second reason is that the connectivity provided by these
advanced networks is expanding at a truly fantastic rate. It is becoming
progressively easier and more cost-effective to connect research and
education communities to each other and to the growing variety of
resources and services to which they contribute and on which they
depend. One specific indicator of this phenomenon is provided by the
growth of the NSFNET (see Figure 1).

[FIGURE 1 NOT AVAILABLE IN ASCII TEXT VERSION]

As of March 31, 1991, 2,501 individual networks, including 757
foreign networks, can be reached through the NSFNET. In the last two
years the total number of individual networks that can be reached by
this means has increased by nearly 700 percent while the number of
foreign networks that can be reached through this network has increased
by over 2,000 percent. No one knows precisely how many individual
computers are interconnected by these networks or how many individual
users are served by those computers, but an educated guess is 200,000
computers and 10 million users. This is an impressive amount of
connectivity--connectivity that is increasing at an equally impressive
rate.

Another view of the simplification and connectivity being offered
by these advanced networks is provided by what they promise for library
functions and interfaces. Figure 2 provides a simplified (believe it or
not) conceptualization of typical library functions and how these
functions interface with a variety of external agencies and actors. For
instance: the diagram shows that: (a) patrons interface with the
library's reference staff and system, the library's catalog and
information resources, and the library's circulation and inter-library
loan staff and systems; and (b) the library's acquisitions staff and
system interface with publishers, brokers, and other information
resources. Figure 3 illustrates that a variety of networking
technologies are already being used to enhance the effectiveness and to
increase the efficiency of these interfaces. For instance, it shows
that: (a) private networks are being used to interface the library's
cataloging staff and system with bibliographic networks such as OCLC and
the Research Libraries Information Network; and (b) the library's
reference staff and system interface with services such as DIALOG and
Lexis using, primarily, commercial networks. A third related figure
(Figure 4, page 33) shows how a contemporary advanced network can
simplify the technological characteristics of these existing interfaces
while increasing the number of connections that exist among the full
range of library functions and between the library functions and the
full range of external agencies and actors.

[FIGURES NOT AVAILABLE IN ASCII TEXT VERSION]

Performance is the third reason why information technologists and
librarians are so excited about these advanced networks. Performance
levels are already mindboggling and they promise to be dumbfounding by
1995, if not sooner. Again, the NSFNET provides an object lesson. In
March 1991 the National Science Foundation transported 7.03 billion
packets that averaged 350 characters of information. This impressive
figure becomes even more so when you consider that it represents around
a 250 percent growth in traffic transported by the NSFNET in the single
year that ended in March 1991, a compound growth rate for the year that
averaged around 20 percent per month. No one knows precisely how much
traffic is transported within but not between the individual networks
that are interconnected by the NSFNET, but most analysts believe that a
ten-to-one ratio is a fair estimate. This estimate implies that in March
1991 alone more than 70 billion packets of information were transported
within the networks that are interconnected by the NSFNET. This is a
staggering level of performance.

One way to try to grasp what these levels of performance mean and
will mean to research and education communities is to pose the question
of how many typewritten pages can be transported at a variety of
illustrative performance levels. Some relatively straight-forward
quantitative assumptions lead to some very interesting results. For
instance, if we assume that there are 200 words on a typical typewritten
page, that each word has ten letters, and that each letter requires 10
bits to encode, then we can conclude that it takes 20,000 bits to encode
a typewritten page. We can convert this "bits per typical typewritten
page" figure into a series of "typical typewritten pages per second"
figures by simply dividing it into a series of "bits per second" figures
for typical contemporary and near-future performance levels.

If we assume that all the network modems, controllers, and circuits
in question operate at perfect efficiency, then: (a) at 2.4 kilo
(thousand) bits per second (kbps), the performance level of most
contemporary personal computer modems and circuits, just over a tenth of
a typewritten page can be transported each second; (b) at 9.6 kbps, the
performance level of the next generation (just now coming into wide use)
of personal computer modems and circuits, just under one-half of a
typewritten page can be transported each second; (c) at 1.5 mega
(million) bits per second (mbps), the performance level of most
contemporary network controllers and circuits, seventy-five typewritten
pages can be transported each second; (d) at 45 mbps, the performance
level of the next generation (just now coming into wide use) of network
controllers and circuits, 2,250 typewritten pages can be transported per
second; and (e) at 500 mbps and 1 giga (billion) bits per second (gbps),
the performance levels of the generation of network controllers and
circuits that will be in commercial use by 1995, 25,000 and 50,000
typewritten pages, respectively, can be transported per second.

The meaning of these performance levels can be made clearer still
by considering the case of my personal library of 2,000 books. If we
assume that the typical book starts life as a 1,000-page typewritten
manuscript, then it would take my personal library only forty seconds to
be transported at 1 gbps. It would take a typical academic library of 1
million books only six hours to be transported at 1 gbps. It would take
a relatively large research library of 5 million books only a day and a
quarter to be transported at 1 gbps. There are many analysts who believe
that it is a much better than fifty-fifty proposition that by 1995 we
will achieve production performance levels of 3 gbps rather than 1 gbps.
This is a dumbfounding technological prospect and no one fully
understands what it will mean for research and education communities.

Where did these performance levels and the advanced networks that
utilize them come from and where will they come from in the future?

Research and education institutions and organizations have played the
most important role to date in building and operating these networks,
and they have played that role by making significant technological
innovations as well as by making significant financial investments. Most
higher education institutions already have a campus network or a plan by
which to obtain one; by 1995 this most certainly will also be true for
the overwhelming majority of research and education institutions and
organizations.

In the U. S. the Federal government, through the Advanced Research
Projects Agency of the Department of the Army, the Department of Energy,
the National Science Foundation, the National Aeronautics and Space
Administration, and quite a few other Federal agencies, has played the
second most important role to date in building and operating these
networks. However, state and local governments, and related regional
undertakings, have recently begun to look to advanced networks to
improve the educational and economic opportunities available to their
citizens and residents, and to enhance the effectiveness and to improve
the efficiency of their many civic administrative functions, such as
vehicle registration, property title documentation, and the like.
Similar roles are being played by governments at all levels in countries
around the world.

Private and commercial enterprises (like IBM and MCI) have played
the third most important role to date in building and operating these
advanced networks. The role of such enterprises will become even more
important during the 1990s as a result of their shift of emphasis from
the analog world of switching telephone circuits to the digital world of
routing datagram packets. It is very important for policy and technology
planners at research and education institutions and organizations to
recognize and strategize the growing importance of the roles that are
played by state and local governments, and related regional
undertakings, on the one hand, and private and commercial enterprises,
on the other.

In this context it is vital to recall that research and education
networks have always been designed to address at least three
requirements fundamental to research and education communities, that
have historically been much less so to private and commercial ones.

First of all, research and education networks strive for horizontal
rather than vertical integration. This is to say that these networks are
built and operated to accommodate the fact that the humanists at two
different institutions or organizations have more in common with each
other than they do with, for example, the scientists at their respective
institutions or organizations. Private and commercial networks, on the
other hand, are usually built to integrate the efforts of a variety of
different actors in a common, vertical value or production chain.

Research and education networks must also account for a wider
degree of technological diversity than must private and commercial
networks. This mostly reflects the wide range of institutions and
organizations that these networks must encompass, but it also manifests
the high degree of innovation that characterizes research and education
communities.

Finally, as a general rule, research and education networks are
used in a greater variety of disciplinary and inter-disciplinary
settings and by a more highly skilled population who are generally
engaged in knowledge creation and use to a much higher degree than are
the users of typical private and commercial networks.

All this will change in the 1990s as research and education
networks begin to support the requirements of populations that have been
typical of private and commercial networks, and vice versa. This
convergence is a widely predicted outcome of the conversion of
industrial economies to information and service ones. As this conversion
progresses, private and commercial enterprises will play an increasingly
important role in building and operating advanced research and education
networks. It is vital that research and education communities not lose
sight of their unique requirements during this necessary transition, and
that they gauge the success of the transition by the genuine passing of
the need for their vigilance in this regard.

Before ending this discussion of advanced networks and turning
attention to the information resources and services that have been and
are being enabled by these advanced networks, I would like to offer four
metaphors for a future that is being created by the march of networking
technology.

We are clearly building and operating an electronic infrastructure
that has the potential scope and scale of the many physical
infrastructures, such as road, water, and sewage systems, for which we
have already mobilized the expertise and found the resources. This new
electronic infrastructure will both stimulate and constrain our
activities and aspirations in the same ways that these other types of
infrastructures have throughout modern history. We will conceptualize
and experience this new infrastructure much as we conceptualize and
experience the interstate highway system of today, the single most
popular metaphor for what these advanced networks will represent to us
some day. However, we will use maps, guidebooks, and other reference
tools to navigate and travel in a new space that is a "virtual" rather
than a "physical" presence in our lives. Until the reality of these
advanced networks measures up to the full potential of this vision,
though, the experience of using them will be rather like trying to drink
from a firehose. For some time information will gush forth from such
networks at much greater rates and in much greater volumes than we will
be able to capture, manipulate, or assimilate. This means that for the
foreseeable future there will continue to be a compelling need for
informational intermediaries, such as librarians and other information
specialists, who will acquire, organize, store, and add value to
information even though it is being distributed and accessed by
electronic rather than printed means. It is even arguable that advanced
networks will increase the need for such intermediaries. I also believe
that it is very important for all of us, be we authors, intermediaries,
or readers, to recognize that we cannot predict all, perhaps not even
most, of the new things and behaviors that will emerge and occur in the
new ecologies of thought and communication that these advanced networks
represent. Accordingly, we need to come to think of ourselves as
managing such ecologies as well as building and maintaining such
infrastructures.

What are "networked information resources and services," and how do they
relate to the mission of improving information distribution and access
using high performance computers and advanced networks to support
research and education communication?

Most research and education networks to date have been built and
operated to provide access to computational resources and to other types
of powerful and expensive scientific and technological instruments.
Supercomputers represent the most important contemporary example of this
type of resource. However, once the first research and education
networks became operational and the uses to which they were actually
being put became a subject of investigation, it was discovered that
there was a second resource that was at least equally important, if not
more important, to the users of such networks: this resource is people.

An analysis of the traffic being transported by the NSFNET makes
this abundantly clear. Nearly 25 percent of this traffic is accounted
for by electronic mail, and some very large portion of the 25 percent of
the traffic accounted for by file exchanges results from people sending
files to each other, rather than from computers sending the results of
computations to their users. These figures compare to the less than 20
percent of the traffic that is accounted for by interactive
computational processes per se. New applications and extensions of
electronic mail are now occurring on a self-sustaining basis. In
particular, the last year has witnessed the explosion of "special
interest discussion group" subscription services and the appearance of
nearly twenty refereed electronic journals.

Library catalogs and campus-wide information systems represent a
third category of networked information resources and services, and it
is this category that accounts for the lion's share of the growth and
excitement in contemporary networking. Library catalogs are already far
and away the most frequently found type of database on the Internet, and
the databases of the Research Libraries Group and OCLC are the most
frequently used "free for service" databases on the Internet. These
early efforts may not be self-sustaining and there is certainly much
more to come than has arrived to date, but it is important to take note
of just how quickly libraries have embraced the potential of advanced
networks and how aggressively they are now seeking new ways to put these
networks to work. Databases of primary research and education materials,
known as "digital libraries," and of secondary materials that provide
reference information about the contents of print collections, as well
as of the contents of digital libraries, are beginning to appear on
research and education networks, and the rate at which they will
continue to appear promises to accelerate exponentially.

High-volume print facilities represent a relatively new fourth
category of networked information resources and services. These
facilities are destined to replace the generation of high-volume
photocopiers currently in use at so many research and education
institutions and organizations. I also believe that they will soon offer
a cost-effective alternative to the laser printers that have become such
a familiar feature of academic and corporate life. These facilities will
be used to print information as soon as a person finds and requests it
at his or her institution or organization. For certain types of
information and users, such "on demand/on site" printing will represent
a vast improvement over the current approach of printing and storing all
information for all users in anticipation of demand. In so far as most
studies estimate that one-third of the cost of conventional printed
research and education materials can be attributed to the inventory
activities and distribution channels for those materials, this new
resource also holds particular promise for reducing the expense and
increasing the responsiveness of acquiring such materials.

The likely impact of these high-volume printing facilities should
not be discounted by the widely felt desire, at least in some quarters,
for a completely electronic information distribution and access system.
These facilities will allow us to experiment with a "just-in-time," in
contrast to the long-established "just-in-case," information
distribution and access system. They will also allow us to
reconceptualize the role of paper. The role of paper in the emerging
just-in-time system is as the most affordable and acceptable interface
by which to access and use the information that is contained in an
expanding number of electronic storehouses. This contrasts markedly to
the role that paper plays in the existing just-in-case system as the
exclusive means by which information is delivered, stored, and used.

These high-volume print facilities, located in copy center
operations such as those of Kinko's, may also provide an effective way
to address the lack of universal access to advanced networks. Dial-up
connection to information resources and services on such networks are
adequate and affordable ways to look for and to find relevant
information in electronic formats, but they are too slow and unreliable
to be used to access such information in any volume. The ability to
route such information to a high-volume print facility that is on the
network, and that is located at a nearby copy center operation, provides
the answer to the question of how to benefit from the low cost of dial-
up connections to advanced networks and from the relatively large
information objects that are found on such networks.

Just over the horizon of contemporary networked information
resources and services can be seen a new generation of such resources
and services that apply artificial intelligence techniques in new and
useful ways. Knowledge robots, or "knowbots," are algorithmic constructs
engineered to wander advanced networks searching for information of
interest to the human being whose interests and requirements they
represent. The term "cybernautics" has recently come into use to refer
to the science and practice of creating and using these network travel
agents and navigational advisors. Intelligent databases are collections
of information that are capable of knowing when and how they grow or are
changed and what the significance of their growth and modification is to
a variety of interested parties with whom they are in regular or even
continuous communication.

Networked information resources and services are interesting in their
own rights, but what can they do to ameliorate some of the pressing
problems that face libraries and their constituencies in contemporary
research and education communities?

One of the most notable of these is the skyrocketing costs of library
materials. While serials expenditures in the 119 members of the
Association of Research Libraries (ARL) increased 53 percent and
monographic expenditures increased 19 percent in the last three years,
during the same period the number of serials titles purchased dropped 1
percent and the number of monographic volumes purchased dropped 16
percent. No matter how you cut these facts they add up to the same
thing: much less information is being obtained for much more money.
Nearly 40 percent of ARL members reduced their rate of acquiring new
monographs by 21 percent or more in the last three years. Clearly, the
attention that has been paid to what is known as the "serials pricing
crisis" needs to be complemented by a heightened level of concern about
what this crisis has done to the pattern of monographic acquisitions in
academic and research libraries.

The size of library collections and, therefore, the amount of space
that is needed to house library collections continues to expand at an
exponential rate as well. One effect of the extraordinary increases in
costs of library materials has been to reduce the rate of acquisition of
new materials and, therefore, to reduce the rate of growth of space
requirements. But this can hardly be put forward as an acceptable way to
manage a library and to address its space needs.

Another pressing concern of academic libraries is the
underutilization of materials once they have been acquired: less than 60
percent of the materials in academic and research libraries ever
circulate, and 80 percent of the materials that do circulate do so
relatively soon after they have been acquired. Too many analysts have
been all too quick to explain this phenomenon by decrying the declining
quality of the literature record. Information cannot be used if it
cannot be found, and better access mechanisms increase levels of use--a
belief that has been repeatedly affirmed by collection use studies
performed both before and after the advent of online library information
systems.

So the use of networked information resources and services
promises to reduce the costs of acquiring library materials, to
stabilize the rate of growth of the space required to house library
materials, and to increase the rate of use of library materials. It is
not yet clear that these specific promises will in fact be realized, but
a great deal of contemporary effort is motivated by the hope that they
will.

Two things are very clear in the extremely complicated and somewhat
theoretical area of the "cost/benefit" performance of networked
information resources and services as compared to their print
equivalents.

First, the transition from card (paper) form catalogs to online
ones may have something to tell us about the transition that we may or
may not now be making from paper form publications to electronic ones.
In my experience, card catalogs collapsed and became unworkable under
the pressure of the information explosion. I propose that something
quite similar is happening now with printed primary research and
education materials--the existing system is collapsing and becoming
unworkable. No matter how difficult it is for us to imagine, the
transition from an exclusively print to a progressively more electronic
information distribution and access system may well be something about
which we have very little choice, and it is certainly something about
which our constituencies may have no choice at all.

Second, I believe that research and education communities, and
particularly their libraries, are beginning to shift toward a "make"
posture and away from a "buy" one as the business strategy by which they
gain access to the information resources and services they need. In
addition to the cry to "take back the rights" that I hear in
contemporary forums devoted to the "serials pricing crisis," I now hear
a new call to "take back the means of production." This new interest in
"self-publishing," both personal and institutional, and in partnership
undertakings that build new networked information resources and services
in not-for-profit and barter settings, is well worth watching and
experimenting with.

Networked information resources and services also promise to
improve access to brittle books that have been preserved on microfilm
and then digitally scanned, to enable library services to be available
around the clock and from any point on the campus network, and for
faculty, students, information technologists, and librarians to work
together to effectively manage the information and knowledge that is
essential to the integrity and success of all research and education
communities.

What do we need to do to get ready to benefit from networked information
resources and services?

Networked information resources and services force us to rethink each
and every one of the design assumptions of most of the current
generation of local library information systems. These assumptions can
be paraphrased by saying that most such systems assume that they are
providing service to "... a smart user using a dumb terminal right
around the corner from a large computer that contains descriptions of
information owned by the library." The problem is that in today's world
we are all dealing with dumb (well, perhaps inexperienced) users who are
using personal computers and workstations located almost anywhere on the
planet to access computers of all sizes to obtain information itself, as
well as to access computers to obtain information about information that
is neither owned nor licensed by the same institution or organization
that owns or licenses the computer. In my experience, designers and
vendors of such systems are only too well aware of how completely their
systems need to be rethought in terms of a "networked information"
rather than a "housed information" architecture. It is the buyers and
funders of such systems who now need to recognize this fact and to
account for it in their strategic plans and, even more important, in
their depreciation schedules.

Redesigning local library information systems is only one of the
things that we need to do to get ready to benefit from networked
information resources and services. In general, research and education
institutions and organizations must focus on improving their readiness
in four key areas: campus network, automated library, skilled and
equipped end users, and hospitable culture.

It is tautological to say that campus networks have to become
ubiquitous, affordable, and responsive for the promise of networked
information resources and services to be realized. It is equally
important to have a vital, evolving library technology program and a
skilled and equipped group of end users. But these three readiness
factors, no matter how necessary, are not sufficient. Efforts directed
at these factors need to be planned and executed in a cultural setting
that is hospitable to their purposes and problems. Promotion and tenure
practices, for instance, need to recognize and reward excellence in
authoring networked information as well as recognizing and rewarding
excellence in authoring printed information. Accreditation and
statistical practices and criteria need to rate libraries on how well
they deliver information as well as on how well they buy and maintain
information. Information technologists and librarians need to work
together to construct a "single information system image" for the
faculty, students, administrators, and other stake-holders who depend so
much on their vision, talent, and energy.

It is also extremely important to recognize that contemporary
efforts devoted to advanced networks provide the opportunity to merge
two quite different and equally powerful research and education
networking traditions. During the 1970s and 1980s librarians were
funding and building the Research Libraries Group and OCLC, arguably the
only integrated, nation-wide applications of networking that the
research and education community has ever successfully made. During the
same period, information technologists were funding and building the
ARPANet, BITNET, and NSFNET, and the global Internet, among other
advanced networks. I believe that what we are about right now is the
leveraging of each tradition to the benefit of the other and to the
benefit of the constituencies that are shared by librarians and
information technologists.

The readiness of the information marketplace must also be improved
in at least five key areas as well: pricing, payment, protection,
regulation, and experimentation. The marketplace does not currently know
how to price networked information and even if it did we would not know
how to pay for that information in all the ways and by all the schemes
we need. The marketplace has not come to agreement on ways and means for
protecting networked information, from unauthorized modification as well
as from misuse and misappropriation. The regulatory framework by which
"conduit" is differentiated from "content" is extremely fragile, having
resulted from a series of ad hoc rather than deliberate decisions. The
result is that some lines of business are not allowed for some
enterprises and some activities are judged to produce "unrelated
business income" for some research and education institutions and
organizations. Finally, experimentation with new networked information
resources and services is too costly and risky and the results are too
anecdotal for all parties involved. We simply must devise a much more
satisfactory system of research, development, and dissemination than the
one we have at present.

The Coalition for Networked Information (founded in March 1990 as
an activity of ARL, CAUSE, and EDUCOM) is particularly devoted to
identifying and addressing such institutional and marketplace readiness
factors. The mission of the Coalition is to promote the creation of and
access to information resources in networked environments to enrich
scholarship and to enhance intellectual productivity. Just over 135
separate institutions and organizations have become sponsors of the
Coalition's activities by joining a Task Force to guide that mission.
The real story of the Coalition's progress, though, is told by the
variety of information, service, and technology providers who have
joined numerous research and educational institutions and quite a few
collaborating professional and scholarly societies in a common program
of work devoted to a shared vision of how the nature of information
management must change through the end of the 20th century and into the
beginning of the 21st.

In closing, I'd like to pose four questions that I believe are
fundamental to realizing the full promise of networked information
resources and services, questions that all concerned parties, not just
information technologists and librarians, can relate to and help to
answer.

First, the technical question: what benefits can be achieved,
really? We have to find a way to spend less time on wishful thinking and
more time on improving the performance of the systems and technologies
that we already have. We must figure out ways to get new value out of
these existing assets. We must also be ready, willing, and able to
change the way we have been doing things to leverage these existing
assets to get more things done faster and without a loss of quality.
However, the major thrust of the technical question is the pressing need
to improve our ability to hold technology accountable to providing real
benefits to real people.

Second, the political question: who will experience these benefits
when using what resources? This question has an economic as well as a
political component, but I believe that the political component of the
two is the much more important and in need of attention. We must figure
out ways to become more concerned than we have been to date about how
access to the benefits of networked information resources and services
is obtained. We also must become better at remembering that diverse user
populations enrich and strengthen the design and performance of
technological systems.

The third question calls attention to the role of institutions like
libraries in consolidating the gains of technological advance. It is the
institutional question: how will these benefits be secured and made
routine as soon as possible? We must figure out ways to refit
institutional and organizational facilities, to reallocate institutional
and organizational budgets, and to reskill relevant institutional and
organizational professionals if we are to succeed at embedding networked
information resources and services into the milieu of research and
education communities.

Finally, we must assure ourselves that what we do contributes to
improving the basic conditions of human existence. We can explore that
concern by asking the human question: why will these benefits contribute
to the quality of life and the inspiration of intellect? Without
applying this test to our activities and aspirations we can never know
whether we are working on the things that can make the greatest
difference in the course of human affairs.

In conclusion, I believe that:

  *  our facility with the technical question will determine whether
networked information resources and services will become sandboxes in
which technophiliacs play with their new and quite expensive toys, or
will instead become as useful as we envision;

  *  our facility with the political question will determine whether
these resources and services will become battlefields on which conflicts
about ends and means reflect differences in opportunities, or will
instead become opportunities available to all who seek to learn and
think;

  *  our facility with the institutional question will determine
whether networked information resources and services will become the
products of new marketplaces in which financial means play a
disproportionately influential role, or will instead become familiar and
trusted features of the libraries of research and education communities;

  *  our facility with the human question will determine whether
networked information resources and services will become esoteric tools
used by limited populations for narrow purposes, or will instead become
"fields of dreams" for which the guiding principle will be: "If we build
them, the users will come."

************************************************************************

Acknowledgements

The author appreciates the permission of Ronald Larsen of the University
of Maryland at College Park to include Figures 2, 3, and 4, and also
appreciates the assistance of Merit Network's NSFNET Project Information
Services Group, particularly Ken Horning, in providing Figure 1 and
other NSFNET statistics.

************************************************************************