Path: bloom-picayune.mit.edu!snorkelwacker.mit.edu!americast.com!americast.com\!americast-post
Newsgroups: americast.ieee
From: americast-post@AmeriCast.Com
Organization: American Cybercasting
Approved: americast-post@AmeriCast.com
Subject: SPECIAL REPORT/ELECTRONIC MAIL - E-mail at work
Date: Wed, 4 Nov 92 10:20:19 EST
Message-ID: <1041.art.1992Nov4.102019@AmeriCast.com>

SPECIAL REPORT/ELECTRONIC MAIL - E-mail at work

Tekla S. Perry, Senior Editor


Electronic mail has removed the barriers of time and place
between engineers collaborating on complex design projects


Recently, 53 engineers at Digital Equipment Corp.--spread across
Massachusetts, Arizona, Colorado, Singapore, and Germany--
collaborated on the design of a new disk drive. Most had never
met and phoned each other rarely.  Yet Digital estimates that
this dispersed group finished its project one year sooner and
needed 40 percent fewer people than would have a team assembled
in one building. Another recent project at DEC, the Alpha chip,
an integrated-circuit design, also made effective use of e-mail
by calling on engineers in Israel, Massachusetts, and Scotland.
     Tandem Computers Inc., Cupertino, Calif., has a similar
tale. In June the company announced a central office switching
system for use with the SS7 telephone network. The 40-some
engineers on that project were scattered among five buildings in
California and two in Texas. The project manager called the lead
engineers at each location only once a week and visited those
sites only a few times a year; yet the scattered designers worked
as a cohesive unit.
     Electronic mail networks that reach virtually every employee
make such geographically distributed work teams commonplace at
DEC, Maynard, Mass., and Tandem as well as at other
engineering companies.  "Every new employee at Tandem gets
an ID card and a mail access code," said David Foley,
network architect there. "And they may have the mail access
several days before the ID card." At DEC, said Peter E.
Brown, corporate telecommunications manager, "the first
thing we all do when we get to work in the morning is check
our electronic mail." Companies that have come to rely on
electronic mail have huge international networks. Palo Alto,
Calif.-based Hewlett-Packard Co. has a network of 94 000
mailboxes (some dedicated to groups, others to individual
subscribers), which delivers over 350 million messages every year
to its 90 000 employees. DEC's network sends 110 000 employees
around 50 million messages annually between sites (intrasite mail
is not included in this tally).
     Tandem's net links 11 000 employees and delivers about 60
million messages a year. At Sun Microsystems Inc., Mountain View,
Calif., the network has 12 500 users.
     These systems support different flavors of electronic mail.
Private e-mail is person to person, and can comprise computer-
aided design files and digitized images as well as text. Wide-
distribution e-mail originates with one person but is sent
to many and replaces interoffice memos. Electronic conferencing
allows users to identify topics--often technical, but in some
cases with as little connection to work as movie reviews or
softball scheduling. Anyone may review a running transcript of a
conference and append a comment.
     Engineering managers interviewed by IEEE Spectrum, an
admittedly unscientific sample, typically receive about 25
messages daily; engineers receive more or less, depending on the
intensity of projects they are working on and how many interest
group s they belong to.
     So far, though engineering organizations like these large
corporations are at the leading edge of e-mail usage, others,
particularly small companies, have decided that they have
little need for electronic mail.
     Electronic mail is also in less use in some fields, noted
Rob Kling, professor of information and computer science at the
University of California at Irvine. For example, aerospace
engineers are seldom network users.
     However, engineers and scientists are pioneers in electronic
mail use, and professionals in other industries are following
suit in growing numbers: at the Fortune 2000 corporations, some
11.7 million workers are e-mail users, and that figure is
expected to grow to 27 million by 1995, according to a recent
study by the Electronic Mail Association, Arlington, Va.
     COMPETITIVE EDGE. The advantages of electronic mail are
many. Besides the obvious boon of avoiding telephone tag and time
zone dissonance, e-mail gives companies unprecedented
flexibility. Managers can assemble engineering teams by tapping
the best people for the project without concern for their
location, then disband them as soon as the project is done.
Electronic mail also eliminates the need for stressful transfers
or expensive temporary assignments.
     In fact, some companies have discovered that the time-
honored method of assembling a multifunctional team either from
employees at one site or by means of transfers or regular
meetings is no longer practical. According to a study conducted
last year by David Cedrone, DEC's corporate voice and video
manager, and Edward McDonough, a professor from Northeastern
University in Boston: "As projects become increasingly complex
and greater numbers of uniquely skilled people are needed on
project teams, more and more time will be needed to gather them
together in a common location."
     This flexibility may let companies operate with a smaller
workforce--a specialist, say, may work part-time with two project
teams many hundreds of kilometers away. Also, interdisciplinary
teams may crop up more--a group might not need a full-time
cabinet designer for its project, but would jump at the
opportunity to bring one onto the team part-time, instead of
tossing a finalized design to the cabinet department. "Many
individuals are required in "bursts,'" Cedrone and McDonough
concluded in their study. "Full-time assignment of individuals
rarely matches the real resource demands of a project."
     Electronic mail can also uncover hidden expertise in a
company. When a design team at a networked company runs into a
problem that stumps all team members, it broadcasts a "does
anybody know" request throughout the network, and suggested
solutions often appear in a matter of hours. James Treybig,
president of Tandem, has said that "a person in Switzerland on
electronic mail can say "Help' to 10 000 people (which a person
cannot do on the telephone).  The next morning he may have 15
answers to the problem, of which 13 are wrong. But he has
answers."
     A study at Tandem by researchers Sara Kiesler, Lee Sproull,
and David Constant of Carnegie Mellon University in Pittsburgh
found that during a six-week period, employees broadcast about
seven questions a day that elicited about eight replies each.
 The great majority had first tried and failed to find solutions
using other sources. Some of these replies went into open reply
files that could be accessed by other employees with similar
questions.
     Another Tandem study done by Tom Finholt, now assistant
professor of organizational psychology at the University of
Michigan, Ann Arbor, found that the reply files were accessed
more than a thousand times a month. The engineers who used this
database most were those farthest from Tandem's California
headquarters.
     For some companies, electronic mail has meant the ability to
tap into expertise that would have been completely inaccessible
without it. Report Cedrone and McDonough in their study: "Pockets
of expertise and specialization exist throughout the world."
     At Sun Microsystems, engineers are collaborating with a team
of 33 former supercomputer designers in Moscow, St. Petersburg,
and Novosibirsk to develop compiler software for Sparc
workstations, to be marketed in the United States, Europe, and
Japan. Because the telephone network in Russia is unreliable, and
most of the Russian designers have difficulty with spoken
English, electronic mail is for many project workers the sole
means of transatlantic communication. It is used to debate
complex technical issues as well as deal with more mundane
problems. (The California engineers helped their Russian
counterparts network their workstations after the Russians e-
mailed a file in PostScript, a printer language, that contained
the floor plan of their office building.)
     Project teams that span time zones may speed development by
working round the clock. Hewlett-Packard has research
laboratories in Bristol, England, and Tokyo, as well as in Palo
Alto, and joint projects are becoming common. "The time
difference means [the overseas] engineers are going home as we
are going to work and vice versa," said David Ricci, director of
research services at Hewlett-Packard Laboratories in Palo Alto.
"So we can leave them a message at the end of our day, and they
can pick up and work where we left off, then hand it back to us
in our morning."
     "Electronic mail gives us about a 30 percent gain in
productivity," estimated David Ditzel, director of advanced
development at Sun.
     In discussing a study of electronic mail usage in a number
of software design organizations, Carnegie Mellon researchers
Sproull and Kiesler write in their book Connections: New Ways of
Working in the Networked Organization: "We discovered a very
high correlation between use of electronic mail and group
productivity."
     Electronic mail also leads to increased communications by
engineers in the same building, even in adjoining offices. The
reason, pointed out Jan Walker, a member of the research staff at
DEC's Cambridge Research Laboratory, is that electronic mail is
not as intrusive as a phone call. It does not interrupt the
recipient, and for the sender, takes less time since he or she
need not run through the social amenities. "Getting eight phone
calls would be disruptive when you are in the midst of something,
but receiving eight e-mail messages is not, and often you can
answer with just a yes or a no," Walker told Spectrum.
     A fringe benefit of communicating electronically is that a
record is kept of all conversations. Meeting "notes" can be
distributed to interested parties outside the core team, new team
members can easily review past discussions and get up to speed,
managers can review the reasoning behind a decision, and a
message received by one person can be forwarded easily to others.
     "We go back quite often and refer to old messages for part
numbers or specs," said Glenn Rankin, a development engineer at
HP. "It doesn't require writing things down and putting them
in a notebook somewhere."
     Finally, perhaps least in importance, though not trivial, e-
mail is a bargain.  According to Hewlett-Packard's calculations,
a two-page electronic mail message between any two HP employees
worldwide averages US $0.22; a letter averages $0.51, and
a fax averages $1.66.
     Studies have shown, however, that e-mail does not reduce
travel very much. Robert Lucky, executive director of research at
AT&T Bell Laboratories, Murray HIll, N.J., noted that e-mail
makes it even easier to go on the road because the traveler doe
s not risk losing touch with his or her colleagues. And engineers
still subscribe to technical journals to track developments in
their fields. In fact, according to research by Starr Roxanne
Hiltz, professor of sociology at Upsala College, East Orange,
N.J., e-mail users tend to read more, rather than less, probably
because associates send e-mail messages referring to published
articles.
     But does e-mail really allow distant colleagues to work as
closely together as they would if they were not geographically
separated? The engineers who use it say it does, though sometimes
it takes a little creativity to communicate clearly. At DEC
recently, engineers were trying to launch the manufacturing in
Augusta, Maine, of a new digital router (a dedicated device for
directing data communications traffic) but were stymied by
defects, so they sent an oscilloscope trace to peers in Clonmel,
Ireland. The Irish engineers responded the next day: you must be
using capacitors from such and such a vendor, replace them and
your problem is solved. It was.
     ENDING ISOLATION. The earliest users of electronic mail were
researchers at universities. Since the Advanced Research Projects
Agency Network (Arpanet) provided the first reliable electronic
highway in 1969, communication over the net has accelerated the
process of research.
     Researchers thousands of kilometers apart meet in on-line
discussion groups devoted to their specialties, and communicate
regularly on the net with peers whom they may never have
met.
     "One of the facts about being an expert in something
and working at a university is that a typical university can't
afford to maintain more than one or two experts in a field
on staff; therefore, experts tend to be spread all across
the globe,"said psychologist John Condry, a professor at Cornell
University in Ithaca, N.Y. "But with e-mail, I talk regularly to
a dozen people who are doing the same research I am,
communicating as effortlessly with people in France, Sweden, and
Israel as next door. By forming nodes of interaction, e-mail may
well speed up the process of science."
     Sociologist Hiltz's research on the use of electronic mail
by scientists confirms this conclusion. "Perhaps [electronic
communication] is like an intellectual lonely minds club," she
writes, and indicates it has probably had its biggest impact on
those at the smaller universities. The scientists in her study
reported more communication with both people in their fields and
those in other disciplines that resulted in productivity gains by
"increasing the stock of ideas and providing leads, references,
and other information."
     "Subjectively experienced effects of the increased
communication with a larger network of scientists include the
ability to get "instant feedback' on ideas and to "kick ideas
around' with others when a piece of work is in its formative
stage," Hiltz writes.
     With e-mail, months or years no longer pass between a
researcher's completion of experiments and the dissemination of
results. Now, scientific papers are "published" on the network,
commented on, and often revised and "republished" several times
before they appear in traditional journals or are delivered at
conferences.
     Sometimes the pace of net vs. journal publication trips over
itself. A paper finally published in a journal in, say, September
may reference a paper to be published in the same journal in
November, since the electronic forms of both papers had already
been widely circulated and seen by both authors.
     Some think e-mail may have dramatically sped the response to
the purported demonstration of cold fusion claimed by Stanley
Pons and Martin Fleischmann in March 1989. Information about
their results was distributed on Bitnet, a worldwide academic
network, before it was published in a journal; comments by other
researchers, from quick responses to detailed reports of attempts
to replicate the experiments, followed on the network and in a
flurry of faxes between universities.
     Writes Frank Close in Too Hot to Handle, an account of the
cold fusion announcement and its debunking: "By Monday afternoon
[the day after cold fusion was announced in a press conference]
information was coming in by phone, fax, and Bitnet. A new
genre in scientific communication was being born. You logged in
at any time and read the latest gossip or hard news and sent in
any insights you had gathered yourself. The "paper'-- really a
set of electrical signals propagating through the air--flashed
round the world like an electronic chain letter."
     According to Richard Petrasso, a principal physicist at the
Massachusetts Institute of Technology (MIT) in Cambridge,
electronic mail discussions backed up by preprints of papers and
errata transmitted by fax "had a direct impact on our work,"
leading to the publication of several papers on the topic shortly
after the cold fusion experiments were announced.
     Nate Lewis, a professor at the California Institute of
Technology in Pasadena, said researchers in his cold fusion group
"read  By forming nodes of interaction, e-mail may well speed up
the process of science the latest gossip" on the networks every
day. "It propagated rumors, but it also kept close track on the
facts."
     A drawback of electronic mail, though, is that the ability
to move fast is not always positive. "Anytime there is a rush,
there is less time to contemplate your results," said Petrasso at
MIT. "A scientist needs time to cogitate about his paper.  W
e all make mistakes, and the increased rapidity in communications
is depriving the scientist of the time to think, and talk to
colleagues, and change things before they are made public. As a
consequence, there will be more mistakes committed by scientists;
maybe because of e-mail the mistakes will be discovered more
quickly by others, but it doesn't make it a happier environment."
     These days, however, said Lynn Conway, associate dean of the
college of engineering at the University of Michigan, Ann Arbor,
it is hard to imagine coordinating university researchers without
e-mail, because their varied class, research, and travel
schedules makes them especially hard to track down.
     ENGINEERING TOOL. Electronic mail entered the engineering
workplace in the late 1960s, when corporate engineering research
organizations began communicating with each other and
universities over the Arpanet. By the mid-1970s, it was already
being viewed by engineers as a powerful tool.
     At the Xerox Palo Alto Research Center (PARC) in the 1970s,
collaboration by means of electronic mail allowed then PARC
researcher Conway and California Institute of Technology
professor Carver Mead to develop their ground-breaking
methodology for structured very large-scale integration (VLSI)
design. At first they exchanged only text messages between Palo
Alto and Pasadena, but soon they were shipping actual
instructions for IC layouts electronically over Xerox Corp.'s
companywide mail system.
     By tapping into the Arpanet, students across the nation
could submit design files to an automated file server at PARC.
The server would extract chip designs from the e-mail messages,
collect designs into manufacturable groups, and convert the data
into the appropriate format for semiconductor mask-making.
     E-mail is particularly important in making crash projects
successful, Conway told Spectrum. "It is a powerful medium for
small group coordination round the clock," she said.
"Because the VLSI design project was intense, e-mail was
suitable."
     E-mail was also used at Xerox PARC in the late
1970s and early 1980s for the development of Interpress, a
printing protocol that has evolved into today's PostScript
language. John Warnock, an Interpress designer and now
chairman and chief executive officer of Adobe Systems Inc.,
Mountain View, Calif., has recalled that Interpress designers
were scattered among Palo Alto and El Segundo in California and
Pittsburgh and Philadelphia in Pennsylvania.
     In the early 1980s, Common Lisp, a computer language for
artificial intelligence, was designed by a group of some 60
people at numerous organizations collaborating over the Arpanet.
The group met just twice in three years, and has indicated that t
he language development would not have been possible without the
Arpanet's e-mail capability.
     DOWN SIDE. Electronic mail has its limitations, nonetheless.
According to Walker of the DEC Cambridge Research Laboratory, "It
is very hard to reach a decision about something that is complex
and multifaceted." Walker told Spectrum that she has participated
in a number of lengthy and deep technical discussions carried on
by means of e-mail, but has found that, in the absence of a
structured, face-to-face meeting, a participant rarely takes
charge, summarizes the data presented, and guides the group
toward a solution. "It is great for collecting information, but
it is tough to reach closure, because people just haven't worked
out the processes yet," she said.
     Sociologist Hiltz noted a similar phenomenon in her study of
scientists--that half the users felt that theoretical
controversies in their fields were clarified by the use of
electronic communication, but none felt it helped resolve
them.
     Engineers who spend too much time reading e-mail lose sight
of the forest for the trees, Conway has noticed. "Your point of
view becomes skewed to the present, and you lose
the ability to manage subtle things that take time to build
and grow," she said.
     "There is no inertia in the system anymore," said Robert H.
Anderson, senior information scientist at the Rand Corp., Santa
Monica, Calif., "and there  There is no inertia in the system
anymore, and there are times when inertia is a good thing are
times when inertia is a good thing."
     For those who are new to electronic mail or who use it
seldom, the lack of nuance in keyboarded comments poses hazards.
Write Anderson and Norman Z. Shapiro in a Rand report:
"Perhaps the most important phenomenon in electronic mail
systems is the likelihood that the recipient will react
negatively or inappropriately in reading material that might well
have been misinterpreted."
     An informal system of "inflecting" e-mail, called "smilies"
has developed, but it does not always register with the
recipient. In an electronic mail debate on the risks of the
computer revolution, published in the Wall Street Journal on
April 6, Mitchell Kapor, founder and former chief executive of
Lotus Development Corp., Cambridge, Mass., commented: "Risks,
what risks? Computers are here to benefit all personkind.:-)"
     But debate participants who did not notice the smilie :-)
got into a heated debate over his remark. It was interrupted only
when Kapor revisited the conversation three days later and said,
"The typographic glyph :-) which I included at the end of
my comment is the on-line equivalent of an ironic or sarcastic
tone of voice. It is intended to convey that the writer really
means the opposite of the literal meaning of what preceded???What
I was saying was that there are risks in computers."
     Even when understood, smilies do not eliminate another
hazard of e-mail communication--the temptation to "flame." With
e-mail it is all too easy to dash off an angry and ill-considered
reply to a message--much easier than when responding orally or
 in a formal letter.
     Some electronic users are not aware or forget that "private"
e-mail is not really private; a record of the messages exists on
a file server somewhere, is sometimes archived onto tape and
stored for years, and can be retrieved by others in a company
(Federal law prohibits outsiders from snooping into electronic
communications, but employers can tap in at will. Other legal
protection surrounding e-mail are hazy and are currently the
subject of much debate.)
     The good news is that important mail accidentally "deleted"
can be retrieved. The bad news is private mail may be misused;
people have been fired because of the content of supposedly
private messages. In Colorado City, city council members
discovered to their dismay that the mayor was able to thwart them
on certain issues because he had been regularly reading their
"private" communications. And the reappearance of e-mail memos
sent and apparently deleted by White House officials was key
evidence in the Iran Contra hearings. [Security concerns of
computer communications were addressed in the August issue of
Spectrum, "Data Security," pp. 18-34.]
     For its first decades, only a relatively small research
community had access to electronic mail. But in the past five
years or so, it has become widespread in the biggest
corporations, and it will be some more years before it reaches a
more general audience. As a result, much electronic mail to date
has been of some value to the recipient. Nonetheless junk mail
has begun to proliferate, and may threaten e-mail's usefulness.
To combat the threat, several electronic-mail systems have
introduced "filters" in various versions.
     Some filters sort mail into topic folders defined by the
user, so that less important mail may be collected and handled at
leisure. Others act as "bozo" filters, eliminating messages from
certain sources, giving those from key sources high priority, or
forwarding some to other recipients for handling. Another type
filters out any message sent to more than, say, 20 people,
rationalizing that messages with broad distributions either are
of little account or will be heard through other means.
     Filters, however, are not popular. They prevent people from
getting the unexpected message--the new contact, the new
information--that is sometimes the most important by-product
of electronic mail.
     Even nonjunk mail is sometimes overwhelming--25 messages a
day can be reviewed and handled in 20 minutes, perhaps, but the
several hundreds of messages waiting after a two-week vacation
are daunting, keeping some engineers and managers from ever
really taking a break; they check their electronic mail daily, no
matter where they are or what they are doing.
     A solution suggested by sociologist Hiltz in her 1985 book
Online Communities--but apparently not widely implemented--is a
self-destruct capability: senders can tag messages with the last
date of their usefulness, after which they erase themselves. For
example, Zmail, a mail system used on Symbolics computers, asks
for an expiration date in the header of a message, and users
choose whether or not to have expired messages automatically
deleted.
     RESTRUCTURING THE ORGANIZATION. By its very nature,
electronic mail blasts aside typical corporate hierarchies
because the messages are undifferentiated--there is no fancy
letterhead or secretary to place a call and ask the person called
to hold for president so-and-so. In an electronic list of
messages, a message from a summer intern looks just as important
as one from the company president.
     Electronic mail, said Lucky at AT&T Bell Laboratories, "has
produced a new social fabric for the R&D community that cuts
across corporations and the hierarchy of organizations, creating
a new kind of accessibility. It is easier to send e-mail to very
important people, people whom you would never consider writing or
calling."
     At Tandem, anybody can, and many do, send messages to
president Treybig, an open access policy the company believes is
one reason for its success.
     Electronic mail also eliminates cues about age, gender,
race, and appearance: people are judged only by the value of
their ideas, so all ideas can get an equal hearing. Carnegie
Mellon researcher Kiesler writes: "When communication lacks
dynamic personal information, people focus their attention on the
message rather than on each other."
     This may lead to better decision-making. Kiesler and
Sproull, also at Carnegie Mellon, concluded in their studies that
the best solution may not arise from face-to-face group
discussions if it is suggested by a low-status person; electronic
meetings may find better answers.
     And for engineers, Lucky said, e-mail may be a better medium
for communication than voice. "I'm shy about talking to people on
the telephone that I haven't met in person," Lucky told Spectrum.
"I would rather deal with a computer. It's easier to think with
my fingers, and I don't get tongue-tied."
     Electronic mail reduces the tendency of the more outspoken
personalities to dominate discussions, and therefore allows more
diverse ideas to emerge. As Kiesler and Sproull put it, research
laboratories have leapers, who think quickly on their feet
 and love debate, and plodders, who like to work through
implications of ideas in detail before sharing their analysis.
With electronic mail, both types of scientists get their ideas
heard; without it, the leapers dominate.
     The ability to communicate across the hierarchy, broadcast
to peers in various places in an organization, and form ad hoc
communities to solve problems will reduce the need for middle
management and make corporations more efficient, said Rand
Corp.'s Anderson. "Corporations can be run with thousands of
employees less than before," he told Spectrum, "and we are
already seeing some layoffs result."
     The effect is positive, Anderson indicated, because the
talents of individual employees will be able to develop more
fully than if they were locked in a traditional physically co-
located work group. "A lot of people will have to find new
employment," he said, "but somebody always gets hurt in
revolutions, and this is a revolution."

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

         Meta-matrices

"Uncontrollable" is the best way to describe the growth of big
networks, said Vinton Cerf, president of the recently formed
Internet Society, and a computer scientist at the Corporation for
National Research Initiatives located in Reston, Va. No on
e really knows just how extensive many of these decentralized
computing webs of cyberspace are. The indicators are, however,
that the proliferation is occurring exponentially.
     A half-dozen varieties of wide-area networks exist. Besides
the many commercial offerings such as Prodigy and CompuServe [To
probe further, p. 33], there are the exchanges for science,
technology, and education. The Internet is the largest
all-purpose global meta-network supported by governments, while
FidoNet represents another class of connection, the kind formed
spontaneously by individuals without much investment.
     INTERNATIONAL NET. The Internet traces its origin to
Arpanet, a computer science experiment set up by the Pentagon's
Advanced Research Projects Agency in the late 1960s. In the next
decade, Arpanet's growth plus a confluence of terrestrial and
satellite switching technologies and the development of local-
area networks set the stage for wide-area interlinked computer
networks.
     An additional influence came from the U.S. Department of
Defense, which in 1978 endorsed the Transmission Control
Protocol/Internet Protocol (TCP/IP) as a data communications
standard. Devised in part by Cerf, then at California's Stanford
University, TCP/IP in 1983 was made a requirement on Arpanet and
Milnet, a Government military network, by the Defense
Communications Agency.
     After that, companies responded by making TCP/IP-compatible
products, like routers and modems. "That was the beginning of the
explosive period of growth," Cerf told IEEE Spectrum.
     The Internet protocol suite, now widely accepted
internationally, is designed for decentralized use and to link
heterogeneous systems.  The collection of networks that share
this protocol is known as the Internet.
     Once hooked on to the Internet, most users pay no more to
send 1000 messages to Tokyo than to send 10 to Boston.  Lower-
echelon users are generally charged a one-time installation fee
of several thousand dollars for a dedicated phone line and
telecommunications gear. Cyndi Mills, manager of NSF Network
Services (NSFnet), Cambridge, Mass., and head of the Internet
Engineering Task Force's accounting working group, said rates
vary from $25 to thousands of dollars a month, according to how
much ban dwidth is rented and the size of the organization.
     FLOWING TRAFFIC. The Internet is analogous to a highway
system, with dedicated communications links--copper and glass
fiber cables as well as satellites--functioning as the concrete
and asphalt. Leased phone lines of 56 kilobits to 1.5 megabits
per second often serve as the on-ramps, connecting to regional
networks. The capacity of the T1 highways is 1.5 Mb/s; that of
the T3 routes is 45 Mb/s. These latter are currently being
installed by a nonprofit joint venture by IBM, MCI, and Merit
Network called Advanced Network and Services Inc., located in
Elmsford, N.Y.
     Breaking messages into various sizes of packets, which are
then sent along optimum routes, helps to keep traffic flowing and
to make efficient use of the expensive high-capacity links.
Though gigabit-per-second superhighways are still in the research
stage, they will be essential for bandwidth-intensive imagery.
(Recently the MIME Internet standard for sending multimedia e-
mail was completed. In July, during a technical conference, voice
and images using the MIME standard were sent over the Internet to
several countries.)
     About 17 000 networks now plug into the Internet, and its
users number in the millions. According to a quarterly survey
done in July by Mark K. Lottor, a consultant for the network
information systems center of SRI International, Menlo Park,
Calif., the Internet has 992 000 host computers, up by 100 000
since April.
     Data on the NSFnet from Merit Network Inc., Ann Arbor,
Mich., show how the largest backbone of the Internet is used. In
June, 15.7 billion packets were transmitted on NSFnet, more than
double the number of June 1991 and five times that of June 199
0. File exchanges accounted for 31 percent of the usage;
electronic mail, for 21 percent; and interactive computing such
as telnet (where a user logs in remotely to operate, say, a
supercomputer), for 13 percent.
     NSFnet is truly international. Over and above the 3898
networks in the United States linked to it, as of July 1992
NSFnet had a total of 30 networks in Brazil, 3 in Estonia, 10 in
Poland, 119 in Japan, 187 in the United Kingdom, 287 in Germany,
243 in France, and 253 in Canada.
     Like postal and highway systems, Internet receives subsidies
from assorted government agencies for research, installation,
maintenance, and service help. Consequently, the cost of a single
message is "virtually impossible to figure out," said Larry
Landweber, vice president of the Internet Society. Institutions
share costs by renting capacity, unlike on Milnet, where users
are charged according to the number of packets sent.
     In the United States, Government-funded backbones such as
NSFnet are ostensibly used for educational and research purposes
only. There are also commercial backbones on the Internet, which
can sell software or services and offer games with the meter
running.
     CALLING FIDONET. Unlike the Internet, FidoNet is a
telephone-based relay network, requiring people to make calls
using existing public phone lines, ideally at regular intervals,
to forward e-mail. FidoNet is acknowledged to be in more than 60
countries, including the United States.
     Since it does not require much infrastructure, it is easily
installed and is therefore common in developing countries, noted
Landweber, a professor of computer sciences and a specialist in
international networking at the University of Wisconsin in
 Madison. Because the FidoNet is financed almost entirely by
individuals, reducing modem-telephone time has been the priority
of the protocols, which now use Zmodem-based transports.
     Since November 1991, an experimental system that uses
Internet to exchange mail and news between Europe and North
America has saved FidoNet operators thousands of dollars a month,
according to Randy Bush, Pacific Systems Group, Portland, Ore.
FidoNet has tens of thousands of public and private nodes and
more than a million users, Bush estimated. The daily volume of
compressed electronic news on FidoNet is about 5 Mb.
     Another problem with e-mail is the maze of address names.
Some business cards display three or more e-mail addresses. An
Internet address might look like: jadam@ieee.org; but to pass to
FidoNet from the Internet, the address might be:
john.adam@p0.f42.n105.z1.fidonet.org. Still other e-mail
addresses use ! or ?  symbols and are even longer.
     In contrast, the labyrinth of protocols causes no big
hassles for electronic mail. It is relatively easy at a gateway
between networks to translate the header from one recognized mail
application protocol to another (such as the Open System
Interconnection's X.400 to that used on the Internet, SMPT and
RFC822).
     "We will always see multiple standards out there," said
Cerf, because of the installed base, different priorities, and
new technology. Even proprietary protocols such as Appletalk
and Decnet may be encapsulated and sent across the Internet.
     The long-awaited privacy-enhanced electronic mail is just
becoming available to Internet users. As companies use
networks to send proprietary information, privacy and
authenticity become essential. But U.S. export restrictions
on certain cryptography techniques may create special islands of
users.
     Cerf also acknowledged that the system is not friendly to
neophytes but rather "is designed and used by people who are
comfortable in the arcane world of software." However, this
culture is changing, he noted. The growing number of commercial
users will surely not put up with difficult user interfaces and,
just as important, will spend a lot of money to spur the market
for easy-to-use systems

Copyright 1992, IEEE Spectrum.

For more information, send-email to American Cybercasting Corporation
(usa@AmeriCast.COM)