Network Working Group M. Kapor
Request for Comments: 1259 Electronic Frontier Foundation
September 1991
Building The Open Road:
The NREN As Test-Bed For The National Public Network
Status of this Memo
This memo provides information for the Internet community. It does
not specify an Internet standard. Distribution of this memo is
unlimited.
Introduction
A debate has begun about the future of America's communications
infrastructure. At stake is the future of the web of information
links organically evolving from computer and telephone systems. By
the end of the next decade, these links will connect nearly all homes
and businesses in the U.S. They will serve as the main channels for
commerce, learning, education, and entertainment in our society. The
new information infrastructure will not be created in a single step:
neither by a massive infusion of public funds, nor with the private
capital of a few tycoons, such as those who built the railroads.
Rather the national, public broadband digital network will emerge
from the "convergence" of the public telephone network, the cable
television distribution system, and other networks such as the
Internet.
The United States Congress is now taking a critical step toward what
I call the National Public Network, with its authorization of the
National Research and Education Network (NREN, pronounced "en-ren").
Not only will the NREN meet the computer and communication needs of
scientists, researchers, and educators, but also, if properly
implemented, it could demonstrate how a broadband network can be used
in the future. As policy makers debate the role of the public
telephone and other existing information networks in the nation's
information infrastructure, the NREN can serve as a working test-bed
for new technologies, applications, and governing policies that will
ultimately shape the larger national network. Congress has indicated
its intention that the NREN
would provide American researchers and educators with the computer
and information resources they need, while demonstrating how
advanced computer, high speed networks, and electronic databases
can improve the national information infrastructure for use by all
Kapor [Page 1]
RFC 1259 Building The Open Road September 1991
Americans. (1)
As currently envisioned, the NREN
would connect more than one million people at more than one
thousand colleges, universities, laboratories, and hospitals
throughout the country, giving them access to computing power and
information -- resources unavailable anywhere today -- and making
possible the rapid proliferation of a truly nationwide, ubiquitous
network... (2)
The combined demand of these users would develop innovative new
services and further stimulate demand for existing network
applications. Library information services, for example, have
already grown dramatically on the NREN's predecessor, the Internet,
because the
enhanced connectivity permits scholars and researchers to
communicate in new and different ways.... Clearly, to be
successful, effective, and of use to the academic and research
communities, the NREN must be designed to nurture and accommodate
both the current as will as future yet unknown uses of valuable
information resources. (3)
So as the NREN implementation process progresses, it is vital that
the opportunities to stimulate innovative new information
technologies be kept in mind, along with the specific needs of the
mission agencies which will come to depend on the network.
Far from evolving into the whole of the National Public Network
itself, the NREN is best thought of as a prototype for the NPN, which
will emerge over time from the phone system, cable television, and
many computer networks. But the NREN is a growth site which, unlike
privately controlled systems, can be consciously shaped to meet
public needs. For a wide variety of services, some of which might
not be commercially viable at the outset, the NREN can
provide selective access that proves feasibility and leads to the
creation of a commercial infrastructure that can support universal
services.... If we fully focus on ...[current] goals and work our
way through a multitude of technical and operational issues in the
process, then the success of the NREN will fully support its
extension to broader uses in the years to follow. (4)
In order to function as an effective test-bed, one that promotes
broad access to a range of innovative, developing services, the NREN
must be built so that it is easy for developers to offer new kinds of
applications, and is accessible to a diversity of users. For
Kapor [Page 2]
RFC 1259 Building The Open Road September 1991
example, to encourage the development of creative, advanced library
services, it must be easy for libraries to open their data bases to
users all across the network. And if these library services are to
flourish through the NREN, then the services must be available to
researchers and students all over the country, through a variety of
channels. Though the NREN itself is intended to meet the
supercomputing and networking needs of the government-financed
research community, Congress has wisely recognized that it can also
function as a channel for delivery of a wide range of privately-
developed information services. To
encourage use of the Network by commercial information service
providers, where technically feasible, the Network shall have
accounting mechanisms which allow, where appropriate, users or
groups of users to be charged for their usage of copyrighted
materials over the Network. (5)
Congress can create an environment that stimulates information
entrepreneurship by mandating that the NREN rely on open technical
standards whose specifications are not controlled by any private
parties and which are freely available for all to use. Such non-
proprietary standards will ensure that different parts of the network
built and operated by independent parties, will all work together
properly. By employing widely-used, non-proprietary standards the
NREN will make it easy for new information providers to offer their
wares on the network. The market will snowball: as more services are
offered, more users will be attracted, who will increase overall
demand. The NREN will also be a test-bed for development and
experimentation with new networking standards that facilitate even
broader, more efficient interconnection than now possible on the
Internet. But throughout the stages of the NREN, all concerned
should be sure that these functionalities are fostered.
The NREN design and construction process is complex and will have
significant effects on future communications infrastructure design:
Building the NREN has frequently been described as akin to
building a house, with various layers of the network architecture
compared to parts of the house. In an expanded view of this
analogy, planning the NII [national information infrastructure] is
like designing a large, urban city.
The NREN is a big new subdivision on the edge of the metropolis,
reserved for researchers and educators. It is going to be built
first and is going to look lonely out there in the middle of the
pasture for a while. But the city will grow up around it in time,
and as construction proceeds, the misadventures encountered in the
NREN subdivision will not have to be repeated in others. And
Kapor [Page 3]
RFC 1259 Building The Open Road September 1991
there will be many house designs, not just those the NREN families
are comfortable with.... The lessons we learn today in building
the NREN will be used tomorrow in building the NII. (6)
The coming implementation and design of the NREN offers us a critical
opportunity to shape a small but important part of the National
Public Network.
VISIONS
At its best, the National Public Network would be the source of
immense social benefits. As a means of increasing social
cohesiveness, while retaining the diversity that is an American
strength, the network could help revitalize this country's business
and culture. As Senator Gore has said, the new national network that
is emerging is one of the "smokestack industries of the information
age." (7) It will increase the amount of individual participation in
common enterprise and politics. It could also galvanize a new set of
relationships -- business and personal -- between Americans and the
rest of the world.
The names and particular visions of the emerging information
infrastructure vary from one observer to another. (8) Senator Gore
calls it the "National Information Superhighway." Prof. Michael
Dertouzos imagines a "National Information Infrastructure [which] ...
would be a common resource of computer-communications services, as
easy to use and as important as the telephone network, the electric
power grid, and the interstate highways." (9) I call it the National
Public Network (NPN), in recognition of the vital role information
technology has come to play in public life and all that it has to
offer, if designed with the public good in mind.
To what uses can we reasonably expect people to use a National Public
Network? We don't know. Indeed, we probably can't know -- the users
of the network will surprise us. That's exactly what happened in the
early days of the personal computer industry, when the first
spreadsheet program, VisiCalc, spurred sales of the Apple II
computer. Apple founders Steve Jobs and Steve Wozniak did not design
the spreadsheet; they did not even conceive of it. They created a
platform which allowed someone else to bring the spreadsheet into
being, and all the parties profited as a result, including the users.
Based on today's systems, however, we can make a few educated guesses
about the National Public Network. We know that, like the telephone,
it will serve both business and recreation needs, as well as offering
crucial community services. Messaging will be popular: time and time
again, from the ARPAnet to Prodigy, people have surprised network
planners with their eagerness to exchange mail. "Mail" will not just
Kapor [Page 4]
RFC 1259 Building The Open Road September 1991
mean voice and text, but also pictures and video -- no doubt with
many new variations. One might imagine two people poring over a
manuscript from opposite ends of the country, marking it up
simultaneously and seeing each others' markings appear on the screen.
We know from past demand on the Internet and commercial personal
computer networks that the network will be used for electronic
assembly -- virtual town halls, village greens, and coffee houses,
again taking place not just through shared text (as in today's
computer networks), but with multi-media transmissions, including
images, voice, and video. Unlike the telephone, this network will
also be a publications medium, distributing electronic newsletters,
video clips, and interpreted reports. (10)
We can speculate but cannot be sure about novel uses of the network.
An information marketplace will include electronic invoicing,
billing, listing, brokering, advertising, comparison-shopping, and
matchmaking of various kinds. "Video on demand" will not just mean
ordering current movies, as if they were spooling down from the local
videotape store, but opening floodgates to vast new amounts of
independent work, with high quality thanks to plummeting prices of
professional-quality desktop video editors. Customers will grow used
to dialing up two-minute demos of homemade videos before ordering the
full program and storing it on their own blank tape.
There will be other important uses of the network as a simulation
medium for experiences which are impossible to obtain in the mundane
world. If scientists want to explore the surface of a molecule,
they'll do it in simulated form, using wrap-around three-dimensional
animated graphics that create a convincing illusion of being in a
physical place. This visualization of objects from molecules to
galaxies is already becoming an extraordinarily powerful scientific
tool. Networks will amplify this power to the point that these
simulation tools take their place as fundamental scientific apparatus
alongside microscopes and telescopes. Less exotically, a consumer or
student might walk around the inside of a working internal combustion
engine -- without getting burned.
Perhaps the most significant change the National Public Network will
afford us is a new mode of building communities -- as the telephone,
radio, and television did. People often think of electronic
"communities" as far-flung communities of interest between followers
of a particular discipline. But we are learning, through examples
like the PEN system in Santa Monica and the Old Colorado City system
in Colorado Springs, that digital media can serve as a local nexus,
an evanescent meeting-ground, that adds levels of texture to
relationships between people in a particular locale. As Jerry Berman
of the ACLU Information Technology Project has said:
Kapor [Page 5]
RFC 1259 Building The Open Road September 1991
Computer and communications technologies are transforming speech
into electronic formats and shifting the locus of the marketplace
of ideas from traditional public places to the new electronic
public forums established over telephone, cable, and related
electronic communications networks. (11)
To both local and long-distance communities, accessible digital
communications will be increasingly important; by the end of this
decade, the "body politic," the "body social," and the "body
commercial" of this country will depend on a nervous system of
fiber-optic lines and computer switches.
But whatever details of the vision and names gives to the final
product, a network that is responsive to a wide spectrum of human
needs will not evolve by default. Just as it is necessary for an
architect to know how to make a home suitable for human habitation,
it is necessary to consider how humans will actually use the network
in order to design it.
In that spirit, I offer a set of recommendations for the evolution of
the National Public Network. I first encountered many of the
fundamental ideas underlying these proposals in the computer
networking community. Some of these recommendations address
immediate concerns; others are more long-term. There is a focus on
the role of public access and commercial experiments in the NREN,
which complement its research and education mission. The
recommendations are organized here according to the main needs which
they will serve: first ensuring that the design and use of the
network remains open to diversity, second, safeguarding the freedom
of users. The ultimate goal is to develop a habitable, usable and
sustainable system -- a nation of electronic neighborhoods that
people will feel comfortable living within.
I. Encourage Competition Among Carriers
In the context of the NREN, act now to create a level and competitive
playing field for private network carriers, (whether for-profit or
not-for-profit) to compete. Do not give a monopoly to any carrier.
The growing network must be a site where competitive energy produces
innovation for the public benefit, not the refuge of monopolists.
The post-divestiture phone system offers us a valuable lesson: a
telecommunications network can be managed effectively by separate
companies -- even including bitter opponents like AT&T and MCI -- as
long as they can connect equitably and seamlessly from the user's
standpoint. The deregulated telecommunications system may not work
perfectly and may produce too much litigation, but it does work. We
Kapor [Page 6]
RFC 1259 Building The Open Road September 1991
should never go back to any monopoly arrangement like the pre-
divestiture AT&T which held back market-driven innovation in
telecommunications for half a century. Given the interconnection
technology now available, we should never again have to accept the
argument that we have to sacrifice interoperability for efficiency,
reliability, or easy-of-use.
Similarly, the NREN, and later the National Public Network, must be
allowed to grow without being dominated by any single company.
Contracting requirements in the current legislation advance this
goal.
The Network shall be established in a manner which fosters and
maintains competition within the telecommunications industry and
promotes the development of interconnected high-speed data
networks by the private sector. (12)
Absent a truly competitive environment, a dominant carrier might use
its privileged access to stifle competitors unfairly: "Use our local
service to connect to our undersea international links, without the
$3 surcharge we tack on for other carriers." The greatest danger is
"balkanization" -- in which the net is broken up into islands, each
developing separately, without enough interconnecting bridges to
satisfy users' desires for universal connectivity. Strong
interoperability requirements and adherence to standards must be
built into the design of the NREN from the outset. (13)
After 1992, private companies will manage an ever-greater share of
the NREN cables and switches. The NSF should use both carrot and
stick to encourage as much interconnection as possible. For example,
the NSF could make funding to NREN backbone carriers contingent on
participation in an internetwork exchange agreement that would serve
as a framework for a standards-based environment. As the NREN is
implemented, some formal affirmation of fair access is needed --
ideally by an "Internet Exchange Association" formed to settle common
rules and standards. (Their efforts, if strong enough, could
forestall a costly, wasteful crazy-quilt of new regulations from the
FCC and 50 State Public Utilities Commissions.) This association
should decide upon a "basket" of standard services -- including
messaging, directories, international connections, access to
information providers, billing, and probably more -- that are
guaranteed for universal interconnection. The Commercial Internet
Exchange (CIX) formed in 1991 by three commercial inter-networking
carriers represents a substantive, initial move in this direction.
Kapor [Page 7]
RFC 1259 Building The Open Road September 1991
II. Create an Open Platform for Innovation
Encourage information entrepreneurship through an open architecture
(non-proprietary) platform, with low barriers to entry for
information providers.
The most valuable contribution of the computer industry in the past
generation is not a machine, but an idea -- the principle of open
architecture. Typically, a hardware company (an Apple or IBM, for
instance) neither designs its own applications software nor requires
licenses of its application vendors. Both practices were the norm in
the mainframe era of computing. Instead, in the personal computer
market, the hardware company creates a "platform" -- a common set of
specifications, published openly so that other, often smaller,
independent firms can develop their own products (like the
spreadsheet program) to work with it. In this way, the host company
takes advantage of the smaller companies' ingenuity and creativity.
Even interfaces rigidly controlled by a single manufacturer, like the
Macintosh, embrace the platform concept. Two years ago, when Apple
began planning the System 7 release of its Macintosh operating
system, one of its first steps was to invite comment from software
companies like Macromind, Aldus, Silicon Beach, and T/Maker. In
substantive, sometimes very argumentative sessions, Apple revealed
the capabilities it planned to these independents, who knew their
customers and needs much better than Apple. One multi-media company,
after arguing that Apple should take a different technical turn,
actually found itself doing the work in a joint project. The most
useful job of Apple's famous "evangelists" is not selling the Mac
specs, but listening to outsiders, and helping Apple itself stay
flexible enough to work with independent innovators effectively.
In the design of the NREN, information entrepreneurship can best be
promoted by building with open standards, and by making the network
attractive to as many service providers and developers as possible.
The standards adopted must meet the needs of a broad range of users,
not just narrow needs of the mission agencies that are responsible
for overseeing the early stages of the NREN. Positive efforts should
be made to encourage the development of experimental commercial
services of all kinds without requiring the negotiation of any
bureaucratic procedures.
In the early stages of development of an industry, low barriers to
entry stimulate competition. They enable a very large initial set of
products for consumers to choose from. Out of these the market will
learn to ignore almost all in order to standardize on a few, such as
a Lotus 1-2-3. The winners will be widely emulated in the next
generation of products, which will in turn generate a more refined
Kapor [Page 8]
RFC 1259 Building The Open Road September 1991
form of marketplace feedback. In this fashion, early chaos evolves
quickly a set of high-demand products and product categories.
This process of market-mediated innovation is best catalyzed by
creating an environment in which it is inexpensive and easy for
entrepreneurs to develop products. The greater the number of
independent enterprises, each of which puts at voluntary risk the
intellectual and economic capital of risk-takers, is the best way to
find out what the market really wants. The businesses which succeed
in this are the ones which will prosper.
It is worthwhile to note that not a single major PC software company
today dates from the mainframe era. Yesterday's garage shop is
today's billion-dollar enterprise. Policies for the NPN should
therefore not only accommodate existing information industry
interests, but anticipate and promote the next generate of
entrepreneurs.
The diverse needs of these many users will create demand for
thousands of information proprietors on the net, just as there are
thousands of producers of personal computer software today and
thousands of publishers of books and magazines. It should be as easy
to provide an information service as to order a business telephone.
Large and small information providers will probably coexist as they
do in book publishing, where the players range from multi-billion-
dollar international conglomerates to firms whose head office is a
kitchen table. They can coexist because everyone has access to
production and distribution facilities -- printing presses,
typography, and the U.S. mails and delivery services -- on a non-
discriminatory basis. In fact, the sub-commercial print publications
are an ecological breeding ground, through which mainstream authors
and editors rise. No one can guarantee when an application as useful
as the spreadsheet will emerge for the NPN (as it did for personal
computers), but open architecture is the best way for it to happen
and let it spread when it does.
The PC revolution was brought about without direct public support.
Entrepreneurs risked their investors' capital for the sake of
opportunity. Some succeeded, but many others lost their entire
investment. This is the way of the marketplace. We should take a
much more cautious attitude about the commitment of public monies.
In the absence of proven demand for new applications, government
should not be spending billions of dollars on the creation of
broadband networks. Neither should telephone companies be allowed to
pass on the costs of the NPN in a way which would raise the rates for
ordinary voice telephone service.
Instead, we should position the NREN to show there is a market for
Kapor [Page 9]
RFC 1259 Building The Open Road September 1991
network applications. The commercial experiments just beginning on
the Internet provides one source of innovation. Deployment of a
national ISDN platform in the next few years represents another
relatively inexpensive seed bed. As such experiments demonstrate
more of a proven demand for public network services, it should be
possible for the private sector to make the investments to build the
broadband NPN using experience from the NREN.
At the same time as the NREN is being debated and developed,
telephone companies continue to push at the limits imposed on them by
the "Modification of Final Judgment" (MFJ) of divestiture, the 1982
anti-trust agreement which split up the Bell system. (14) Under
pressure from the D.C. Court of Appeals, Judge Greene recently lifted
the information services restrictions on the BOCs -- despite the
competitive tension between the telephone companies, cable TV
carriers, and newspapers. Thus, in the next year or so, Congress may
well be forced to define a new set of rules for regulated
telecommunications. (15) Like the AT&T divestiture decision, this
would represent a fundamental shift in national policy with enormous
and unpredictable consequences.
Many consumer and industry groups are concerned that as the MFJ
restrictions are lifted, the RBOCs will come to dominate the design
of the emerging National Public Network, shaping it more to
accommodate their business goals than the public interest. The
Communications Policy Forum, a coalition of public interest and
industry groups, has recently begun to consider what kinds of
safeguards will be needed to maintain a competitive information
services market that allows RBOC participation. The role that the
RBOCs come to play in the nation's telecommunications infrastructure
is, of course, an issue that must be carefully considered on its own.
But in this context, the NREN represents a critical opportunity to
create a model for what a public network has to offer, free from
commercial pressures.
With all of the uncertainty that surrounds the RBOCs entry into the
information services market, we should use the NREN to learn how to
develop a network environment where competitive entry is easy enough
that the RBOCs opportunity to engage in anti-competitive behavior
would be minimized. There is evidence that the RBOCs are resisting
attempts to transform the public telephone system into a truly open
public network (16) notwithstanding the FCCs stated intention do
implement Open Network Architecture. (17) But since the NREN
standards and procedures can be designed away from the dominance of
the RBOCs, a fully open network design is within reach. In this
sense the NREN can be a test-bed for "safeguards" against market
abuse just as it is a test ground for new technical standards and
innovative network applications.
Kapor [Page 10]
RFC 1259 Building The Open Road September 1991
An open platform network model carrier from the NREN to the National
Public Network would actually make some MFJ restrictions less
necessary. Phone companies were originally prohibited from being
information providers because their bottleneck control over the local
exchange hubs gives them an unfair advantage. But on a network in
which the local switch is open to information providers -- because
the platform itself is so rich and well-designed -- creativity and
quality triumph over monopoly power. Instead of restricting
information providers, the National Public Network developers should
encourage the entry of as many new parties as possible. Just as
personal computer companies started in garages and attics, so will
tomorrow's information entrepreneurs, if we give them a chance.
Their prototypes today, small computer networks, electronic
newsletters, and chat lines, are among the most vibrant and
imaginative "publishers" in the world.
III. Encourage Pricing for Universal Access
Everyone agrees in the abstract with universal service -- the idea
that any individual who wishes should be able to connect to a
National Public Network. But that's only a platitude unless
accompanied by an inclusive pricing plan.
The importance of extending universal access to information and
communication resources has been widely recognized:
In light of the possibilities for new service offerings by the
21st century, as well as the growing importance of
telecommunications and information services to US economic and
social development, limiting our concept of universal service to
the narrow provision of basic voice telephone service no longer
services the public interest. Added to universal basic telephone
service should be the broader concept of universal opportunity to
access these new technologies and applications. (18)
The problem of disparate access to information resources has been
recognized in other telecommunications arenas as well. Congressman
Edward Markey (D-Mass.), Chairman of the Subcommittee of
Telecommunications and Finance of the House Energy and Commerce
Committee warns that:
[i]nformation services are beginning to proliferate. The
challenge before us is how to make them available swiftly to the
largest number of Americans at costs which don't divide the
society into information haves and havenots and in a manner which
does not compromise our adherence to the long-cherished principles
of diversity, competition and common carriage. (19)
Kapor [Page 11]
RFC 1259 Building The Open Road September 1991
To address this problem in the long-term, there is legislation now
pending which would broaden the guarantee of universal phone service
to universal access to advanced telecommunications services. Senator
Burns has proposed that the universal service guarantee statement in
the Communications Act of 1934 should be amended to include access to
a nation-wide, advanced, interactive, interoperable, broadband
communications system available to all people, businesses,
services, organizations, and households..." (20)
In the near term, the NREN can serve as a laboratory for testing a
variety of pricing and access schemes in order to determine how best
to bring basic network services to large numbers of users. The NREN
platform should facilitate the offering of fee-based services for
individuals.
Cable TV is one good model: joining a service requires an investment
of $100 for a TV set, which 99% of households already own, about $50
for a cable hookup, and perhaps $15 per month in basic service.
Anything beyond that, like premium movie channels or pay-per-events
is available at extra cost. Similarly, a carrier providing connection
to the mature National Public Network might charge a one-time startup
fee and then a low fixed monthly rate for access to basic services,
which would include a voice telephone capability.
Because regulators are concerned about any telephone service that
might cause the price of basic voice service to rise, they are
unwilling to approve new services which don't immediately recover
their own costs. They are concerned that any deficit will be passed
on to consumers in the form of higher charges for standard services.
As a result, telephone companies tend to be very conservative in
estimating the demand for new services. Prices for new services turn
out to be much higher than what would be required for universal
digital service. This is a kind of catch-22, in which lower prices
won't be set until demand goes up, but demand will never go up if
prices aren't low enough.
Open architecture could help phone companies offer lower rates for
digital services. If opportunities and incentives exist for
information entrepreneurs, they will create the services which will
stimulate demand, increase volume, and create more revenue-generating
traffic for the carriers. In a competitive market, with higher
volumes, lower prices follow.
Kapor [Page 12]
RFC 1259 Building The Open Road September 1991
IV. Make the Network Simple to Use
The ideal means of accessing the NPN will not be a personal computer
as we know it today, but a much simpler, streamlined information
appliance - a hybrid of the telephone and the computer.
"Transparency" is the Holy Grail of software designers. When a
program is perfectly transparent, people forget about the fact that
they are using a computer. The mechanics of the program no longer
intrude on their thoughts. The most successful computer programs are
nearly always transparent: a spreadsheet, for instance, is as self-
evident as a ledger page. Once users grasp a few concepts (like rows,
cells, and formula relationships), they can say to themselves,
"What's in cell A-6?" without feeling that they are using an alien
language.
Personal computer communications, by contrast, are practically
opaque. Users must be aware of baud rates, parity, duplex, and file
transfer protocols -- all of which a reasonably well-designed network
could handle for them. It's as if, every time you wanted to drive to
the store, you had to open up the hood and adjust the sparkplugs. On
most Internet systems, it's even worse; newcomers find themselves
confronting what John Perry Barlow calls a "savage user interface."
Messages bounce, conferencing commands are confusing, headers look
like gibberish, none of it is documented, and nobody seems to care.
The excitement about being part of an extended community quickly
vanishes. On a National Public Network, this invites failure. People
without the time to invest in learning arcane commands would simply
not participate. The network would become needlessly exclusionary.
Part of the NREN goal of "expand[ing] the number of researchers,
educators, and students with ... access to high performance computing
resources" (21) is to make all network applications easy-to-use. As
the experience of the personal computer industry has shown, the only
way to bring information resources to large numbers of people is with
simple, easy-to-learn tools. The NREN can be a place where various
approaches to user-friendly networks are tested and evaluated.
Technically trained people are not troglodytes; they approve of
human-oriented design, even as they manage to use the network today
without it. For years, leaders within the Internet community have
been taking steps to improve ease of use on the network. But the
training of the technical community as a whole has given them little
practice making their digital artifacts appropriate for non-technical
consumption. Nor are they often rewarded for doing so. To a phone
company engineer designing a new high-speed telephone switch, or to a
computer scientist pushing the limits of a data compression
algorithm, the notion of making electronic mail as simple as fax
Kapor [Page 13]
RFC 1259 Building The Open Road September 1991
machine may make sense, but it also feels like someone else's job.
Being technically minded themselves, they feel comfortable with the
specialized software they use and seldom empathize with the neophyte.
The result is a proliferation of arcane, clumsy tools in both
hardware and software, defended by the cognoscenti: "I use the "vi"
editor all the time -- why would anyone have trouble with it?"
If we have the vision and commitment to try this, the transformation
of the network frontier from wilderness to civilization need not
display the brutality of 19th century imperialism. As commercial
opportunities to offer applications and services develop,
entrepreneurs will discover that ease of use sells. The normal,
sometimes slow, play of competitive markets should cause industry to
commit the resources to serve the market by making access more
transparent. But at the start transparency will need deliberate
encouragement -- if only to overcome the inertia of old habits.
V. Develop Standards of Information Presentation
The National Public Network will need an integrated suite of high-
level standards for the exchange of richly formatted and structured
information, whether as text, graphics, sound, or moving images. Use
the NREN as a test-bed for a variety of information presentation and
exchange standards on the road towards an internationally-accepted
set of standards for the National Public Network.
Standards -- the internal language of networks -- are arranged in a
series of layers. The lower levels detail how the networks'
subterranean "wiring" and "plumbing" is managed. Well-developed sets
of lower-level standards such as TCP/IP are in wide use and continue
to be refined and extended, but these alone are not sufficient. The
uppermost layers contain specifications such as how text appears on
the screen and the components of which documents are composed. These
are the kinds of concerns which are directly relevant to users who
wish to communicate. Recently independent efforts to develop high-
level standards for document formats have begun, but these projects
are not yet being integrated into computer networks.
Today, for example, the only common standard for computer text is the
American Standard Code for Information Interchange (ASCII). But
ASCII is inadequate; it ignores fonts, type styles (like boldface and
italics), footnotes, headers, and other formats which people
regularly use. Each word processing program codes these formats
differently, and there is still no intermediary language that can
accommodate all of them. The National Public Network will need such a
language to transcend the visual poverty and monotony of today's
telecommunicated information. It will also need additional standards
beyond what have been developed for message addresses and headers, a
Kapor [Page 14]
RFC 1259 Building The Open Road September 1991
common set of directories (the equivalent of the familiar white pages
and yellow pages directories), common specifications for coding and
decoding images, and standards for other major services.
Congress has provided that the National Institute of Standards and
Technology
shall adopt standards and guidelines ... for the interoperability
of high-performance computers in networks and for common user
interfaces to systems. (22)
As the implementation of the NREN moves forward, we must ensure that
standards development remains both a public and private priority.
Failure to make a commitment to an environment with robust standards
would be "the beginning of a Tower of Babel that we can ill afford."
(23) Since current standards are so inadequate to the demands of
users:
We ... need to endow the NII [National Information Infrastructure]
with a set of widely understood common communication conventions.
Moreover, these conventions should be based on concepts that make
life easier for us humans, rather than for our computer servants.
(24) The development of standards is vital, not just because it
helps ensure an open platform for information providers; it also
makes the network easier to use.
VI. Promote First Amendment Free Expression by
Affirming the Principles of Common Carriage
In a society which relies more and more on electronic communications
media as its primary conduit for expression, full support for First
Amendment values requires extension of the common carrier principle
to all of these new media.
Common carriers are companies which provide conduit services for the
general public. They include railroads, trucking companies, and
airlines as well as telecommunications firms. A communications
common carrier, such as a telephone company is required to provide
its services on a non-discriminatory basis. It has no liability for
the content of any transmission. A telephone company does not concern
itself with the content of a phone call. Neither can it arbitrarily
deny service to anyone. (25) The common carrier's duties have
evolved over hundreds of years in the common law and later statutory
provisions. The rules governing their conduct can be roughly
distilled in a few basic principles. (26) Common carriers have a
duty to:
o provide services in a non-discriminatory manner at a fair
Kapor [Page 15]
RFC 1259 Building The Open Road September 1991
price
o interconnect with other carriers
o provide adequate services
The carriers of the NREN and the National Public Network, whether
telephone companies, cable television companies, or other firms
should be treated in a similar fashion. (27)
Unlike many other countries, our communications infrastructure is
owned by private corporations instead of by the government. Given
Congress' plan to build the NREN with services from privately-owned
carriers, a legislatively-imposed duty of common carriage is
necessary to protect free expression effectively. As Professor Eli
Noam, a former New York State Public Utility Commissioner, explains:
[C]ommon carriage is the practical analog to [the] First Amendment
for electronic speech over privately-owned networks, where the
First Amendment does not necessarily govern directly. (28)
To foster free expression and move the national communications
infrastructure toward a full common carrier regime, all NREN carriers
should be subject to common carriage obligations. Given that the
NREN is designed to promote the development of science, ensuring free
expression is especially important. As on academic said:
I share with many researchers strong belief that much of the power
of science (whether practiced by scientists, engineers, or
clinical researchers) derives from the steadfast commitment to
free and unfettered communication of information and knowledge.
(29)
A telecommunications providers under a common carrier obligation
would have to carry any legal message regardless of its content
whether it is voice, data, images, or sound. For example, if full
common carrier protections were in place for all of the conduit
services offered by the phone company, the terminations of
"controversial" 900 services such as political fundraising would not
be allowed, just as the phone company is now prohibited by the
Communications Act from discriminating in the provision of basic
telephone services. (30) Neither BOCs not IXCs would be allowed to
terminate service because of anticipated harm to their "corporate
image." Though providers of 900 information services did have their
freedom of expression abridged by the BOC/IXC action, First Amendment
protection was not available to them because there was no state
action underlying the termination.
As important as common carriage is to the NPN, it is equally
important that it be implemented in such a way as to avoid sinking
Kapor [Page 16]
RFC 1259 Building The Open Road September 1991
the carriers of these new networks into the same regulatory gridlock
that characterizes much of telecommunications regulation. (31) This
would have a crippling effect of the pace of innovation and is to be
avoided. The controlled environment of the NREN should be taken
advantage of to experiment with various open access, common carriage
rules and enforcement mechanisms to seek regulatory alternatives
other than what has evolved in the public telephone system
Along with promoting free expression, common carriage rules are
important for ensuring a competitive market in information services
on the National Public Network. Our society supports the publication
of many thousands of periodicals and fifty thousand of new books a
year as well as countless brochures, mailings, and other printed
communications. Historically, the expense of producing
professional-quality video programming has been a barrier to the
creation of similar diversity in video. Now the same advances in
computing which created desktop publishing are delivering "desktop
video" which will make it affordable for the smallest business,
agency, or group to create video consumables. The NPN must
incorporate a distribution system of individual choice for the video
explosion.
If the cable company wants to offer a package of program channels, it
should be free to do so. But so should anyone else. There will
continue to be major demand for mass market video entertainment, but
the vision of the NPN should not be limited to this form of content.
Anyone who wishes to offer services to the public should be
guaranteed access over the same fiber optic cable under the principle
of common carriage. From this access will come the entrepreneurial
innovation, and this innovation will create the new forms of media
that exploit the interactive, multimedia capabilities of the NPN.
VII. Protect Personal Privacy
The infrastructure of the NPN should include mechanisms that support
the privacy of information and communication. Building the NREN is
an opportunity to test various data encryption schemes and study
their effectiveness for a variety of communications needs.
Technologies have been developed over the past 20 years which allow
people to safeguard their own privacy. One tool is public-key
encryption, in which an "encoding" key is published freely, while the
"decoder" is kept secret. People who wish to receive encrypted
information give out their public key, which senders use to encrypt
messages. Only the possessor of the private key has the ability to
decipher the meaning.
The privacy of telephone conversations and electronic mail is already
Kapor [Page 17]
RFC 1259 Building The Open Road September 1991
protected by the Electronic Communications Privacy Act. (32) Without
a valid court order, for example, wiretaps of phone conversations are
illegal and private messages are inadmissible in court. Legal
guarantees are not enough, however. Although it is technically
illegal to listen in on cellular telephone conversations, as a
practical matter the law is unenforceable. Imported scanners capable
of receiving all 850 cellular channels are widely available through
the gray market.
Cellular telephone transmissions are carried on radio waves which
travel through the open air. The ECPA provision which makes it
illegal to eavesdrop on a cellular call is the wrong means to the
right end. It sets a dangerous precedent in which, for the first
time, citizens are denied the right to listen to open air
transmissions. In this case, technology provides a better solution.
Privacy protection would be greatly enhanced if public-key encryption
technology were built into the entire range of digital devices, from
telephones to computers. (33) The best way to secure the privacy and
confidentiality Americans say they want is through a combination of
legal and technical methods.
As a system over which not only information but also money will be
transferred, the National Public Network will have enormous potential
for privacy abuse. Some of the dangers could be forestalled now by
building in provisions for security from the beginning.
Conclusion
The chance to influence the shape of a new medium usually arrives
when it is too late: when the medium is frozen in place. Today,
because of the gradual evolution of the National Public Network, and
the unusual awareness people have of its possibilities, there is a
rare opportunity to shape this new medium in the public interest,
without sacrificing diversity or financial return. As with personal
computers, the public interest is also the route to maximum
profitability for nearly all participants in the long run.
The major obstacle is obscurity: technical telecommunications issues
are so complex that people don't realize their importance to human
and political relationships. But be this as it may, these issues are
of paramount importance to the future of this society. Decisions and
plans for the NPN are too crucial to be left to special interests.
If we act now to be inclusive rather than exclusive in the design of
the NPN we can create an open and free electronic community in
America. To fail to do so, and to lose this opportunity, would be
tragic.
Kapor [Page 18]
RFC 1259 Building The Open Road September 1991
End Notes
1. High Performance Computing and Communications Act of 1991, H.R
656, S.272 section 2(6).
2. High-Performance Computing And Communications Act of 1991:
Hearing before the Subcommittee on Science, Technology, and Space of
the Senate Comm. on Commerce, Science, and Transportation, 102nd
Cong., 1st Sess. 1 (1991)(Opening Statement by Senator
Gore)(hereinafter 1991 Senate NREN Hearing).
3. 1991 Senate NREN Hearing 101, 103 (Statement of the Association
of Research Libraries).
4. 1991 Senate NREN Hearing 99 (Statement of Dr. Kenneth M. King,
President, EDUCOM).
5. S.272 (Commerce-Energy compromise) section 102(e)
6. Michael M. Roberts, Positioning the National Research and
Education Network. EDUCOM Magazine 13 (Summer 1991).
7. 1991 Senate NREN Hearing 1 (Opening statement of Sen. Gore).
8. Details of the visions vary in their content and expression.
Senator Gore's bill mandates that federal agencies will serve as
information providers, side by side with commercial services, making
(for instance) government-created information available to the public
over the network. Individuals will gain "access to supercomputers,
computer data bases, other research facilities, and libraries." (Gore
imagines junior high school students dialing in to the Library of
Congress to look up facts for a term paper.) Apple CEO John Sculley
has predicted that "knowledge navigators" will use personal computers
to travel through realms of virtual information via public digital
networks.
Such visions are powerful, but they sometimes seem too much like
sales tools; too vague and overconfident to set direction for
research. People often infer from the Apple's "knowledge navigator"
videotape, for instance, that human-equivalent computer speech
recognition is just around the corner; but in truth, it still
requires fundamental research breakthroughs. Network users will still
need keyboards or pointing devices for many years. Nor will the
network be able (as some have suggested) to translate automatically
between languages. (It will allow translators to work more
effectively, posting their work online.)
9. M. Dertouzos, Building the Information Marketplace, Technology
Kapor [Page 19]
RFC 1259 Building The Open Road September 1991
Review 29, 30 (January 1991).
10. See FCC Hearing on "Networks of the Future" (Testimony of M.
Kapor)(May 1, 1991).
11. J. Berman, Democratizing the Electronic Frontier, Keynote
Address, Third Annual Hawaii Information Network and Technology
Symposium, June 5, 1991.
12. S.272, section 5(d). This section continues: "(1) to the maximum
extent possible, operating facilities need for the Network should be
procured on a competitive basis from private industry; (2) Federal
agencies shall promote research and development leading to deployment
of commercial data communications and telecommunications standards;
and (3) the Network shall be phased into commercial operation as
commercial networks can meet the needs of American researchers and
educators."
13. The distinction between strong support for interoperability and
something less is illustrated in the NREN compromise debate occurring
as this paper is being written. The bill from the Senate Commerce
Committee (S.272) calls for "interoperability among computer
networks," section 701(a)(6)(A), while the compromise currently being
discussed with the Energy Committee adopts a more watered down goal
of "software availability, productivity, capability, portability."
section 701(a)(3)(B).
14. 552 F.Supp 151 (D.D.C. 1982)(Greene, J.). The MFJ restrictions
barred the BOCs from providing long distance services, from
manufacturing telephone equipment, and from providing information
services.
15. The Senate, under the leadership of Sen. Hollings, has just
recently voted to lift the manufacturing restrictions against the
BOCs contained in the MFJ.
16. In The Matter of Advanced Intelligent Network, Petition for
Investigation, filed by Coalition of Open Network Architecture
Parties (November 16, 1990).
17. Amendment of Sections 64.702 of the Commission's Rules and
Regulations, 104 FCC 2d 958 (COMPUTER III), vacated sub nom,
California v. FCC (9th Cir. 1990).
18. NTIA Telecomm 2000 at 79.
19. Committee on Energy and Commerce, Subcommittee on
Telecommunications and Finance, Hearings on Modified Final Judgment,
Kapor [Page 20]
RFC 1259 Building The Open Road September 1991
101st Cong., 1st Sess., 1-2 (May 4, 1989).
20. Communications Competitiveness and Infrastructure Modernization
Act of 1991, S. 1200, Title I, Amending Communications Act section 1,
47 USC 151.
21. S.272, section 2(b)(1)(B).
22. S.272 Commerce-Energy Compromise section 203(a).
23. 1991 Senate NREN Hearing at 32 (Statement of Hon. D. Allan
Bromley, Director, Office of Science and Technology Policy).
24. M. Dertouzos at 31.
25. See 47 USC section 201.
26. See ACLU Information Technology Project, Report to the American
Civil Liberties Board from the Communications Media Committee to
Accompany Proposed Policy Relating To Civil Liberties Goals and
Requirements of the United States Communications Media
Infrastructure. (Draft, July 15, 1991) [hereinafter, ACLU Report].
"Non-discriminatory access to new communications systems must be
guaranteed not simply because it is the economically efficient thing
to do, but more importantly because it is the only way to ensure that
freedom of expression is preserved in the Information Age."
27. Though common carriage principles have historically been applied
to telephone and telegraph systems, the preservation of First
Amendment values of free expression and free press was not the
motivating factor. Professor de Sola Pool notes that telephone and
telegraph systems inherited their common carrier obligations not so
much out of First Amendment concerns, but in order to promote
commerce. The more appropriate model to look to in extending First
Amendment values to new communications technologies is the mails. As
reflected in the post clause, empowering Congress to "establish post
offices and post roads," the Constitutional drafters felt that
creation of a robust postal system was vital in order to ensure free
expression and healthy political debate. As Sen. John Calhoun said
in 1817:
Let us conquer space. It is thus that . . . a citizen of the West
will read the news of Boston still moist from the press. The mail
and the press are the nerves of the body politic.
Non-discriminatory access to the mails has been secured by the
Supreme Court as a vital extension of First Amendment expression. In
a dissent which is now reflective of current law, Justice Holmes
Kapor [Page 21]
RFC 1259 Building The Open Road September 1991
argued that
[t]he United States may give up the Post Office when it sees fit,
but while it carries it on the use of the mails is almost as much
a part of free speech as the right to use our tongues. (Milwaukee
Social Democratic Publishing Co. v. Burleson, 255 US 407 (1921)
(Holmes, J., dissenting)(emphasis added). This principle was
finally affirmed in Hannegan v. Esquire, 327 US 146 (1945) (cited
in de Sola Pool).
See de Sola Pool, Technologies of Freedom 77-107.
28. E. Noam, FCC Hearing "Networks of the Future" (May 1, 1991).
29. 1991 Senate NREN Hearing at 52 (Statement of Donald Langenberg,
Chancellor of the University of Maryland System).
30. 47 USC section 201. Following much controversy about obscene or
indecent dial-a-message services, a number of BOCs and interexchange
carriers (IXCs, ie. MCI, Sprint, etc.) have adopted policies which
limit the kinds of information services for which they will provide
billing and collection services. Recently, some carriers have gone
so far as to refuse to carry the services at all, even if the service
handles its own billing. See ACLU Report.
31. See J. Berman & W. Miller, Communications Policy Overview 14-24,
Communications Policy Forum (April 1991).
32. Electronic Communications Privacy Act of 1986, 18 USC 2510 et
seq. See also J. Berman & J. Goldman, A Federal Right of Information
Privacy: The Need for Reform, Benton Foundation Project on
Communications & Information Policy Options (1989).
33. See Statement In Support Of Communications Privacy, following
1991 Cryptography and Privacy Conference, sponsored by Electronic
Frontier Foundation, Computer Professionals for Social
Responsibility, and RSA Software. (June 10, 1990).
Kapor [Page 22]
RFC 1259 Building The Open Road September 1991
Security Considerations
Security issues are not discussed in this memo.
Author's Address
Mitchell Kapor
Electronic Frontier Foundation
155 Second Street
Cambridge, MA 02142
Phone: (617) 864-1550
EMail: mkapor@eff.org
Kapor [Page 23]