Online Before the Internet, Part
Early Pioneers Tell Their Stories
by Susanne Bjørner • Bjørner &
& Stephanie C. Ardito • Ardito Information &
In 2003, many indeed most people think the word "online" means
the Internet. But there was an "online" before the Internet. In this space
and for succeeding months, we will explore the genesis of the first online
industry, which surfaced in the late 1960s, and served primarily information
scientists; documentation experts; government researchers in educational,
scientific, technical and medical fields; and librarians.
We believe the small but vigorous market that developed around early online
technology paved the way for the phenomenon of the World Wide Web in the mid-1990s,
and that many of the issues intertwined in the growth of early online presaged
those that later exploded with the Web. We don't think those individuals and
companies that ignited information access through the Web learned much from
the development of the first online indeed, we doubt that many, if any,
were particularly aware of it but we do think there are lessons to be
learned from an examination of the growth of the early online industry. This
series presents the stories of some of the important individuals who produced
the first online wave.
Part 1: In the Beginning
The technology of early online was a confluence of systems, databases, and
people. Although they worked together to grow a new industry, the individuals
and companies instrumental in early online also competed with each other. The
two systems that became virtually synonymous with the term "online" in the
1970s and 1980s were Dialog and ORBIT. Both systems owed their early growth
to the persistence of individuals in scientific and technical enterprises.
Dialog came out of the Lockheed Palo Alto Research Laboratory, and ORBIT started
within System Development Corporation, a spin-off of the RAND Corporation.
In the early years of development within these large government contractor
companies, and subsequently during the first years of fledgling commercialization,
both enterprises were headed by strong individuals who, though they did not
work alone, are generally acknowledged to be the fathers of the respective
Therefore, it was an historic occasion in April 2000, when Roger Summit (Dialog)
and Carlos Cuadra (ORBIT) met for a day to talk about the early days of their
rival projects. The meeting took place in the home of Dick Kollin, the creative
genius behind several online innovations (the Pandex database, Magazine Index,
Telebase, and EasyNet). The conversation was initiated and guided (to a small
degree) by Stephanie Ardito and Susanne Bjørner. [Dick Kollin had suffered
a stroke and was assisted in interpretation by his wife Sandy.] It was a day
of reminiscence, revelation, contradiction, corroboration, commiseration, disagreement,
surprise, and laughter. Excerpts from the story that unfolded that day are
offered in this installment. We focus on areas of interaction among these early
online giants, highlighting the evolution and marketing of the systems. Individual
stories regarding the development of their independent companies and
others will follow later in this series.
Key Dates in the Development of Orbit and Dialog
(Despite our attempts, we could not firmly confirm and reconcile
all dates mentioned by participants and the literature. The authors
welcome evidence and recollections from readers to help us verify
events and revise the timeline.)
1960: Roger Summit starts a summer job at Lockheed, assigned to
computer simulation and information retrieval projects.
1963: Summit writes MATICO, a program to print catalog cards.
1964: Lockheed sets up Information Sciences Laboratory and purchases
one of the first IBM 360/30 third-generation computers.
1965: Summit visits Mel Day at NASA in Washington; writes unsolicited
proposal for a demo project on NASA's 300,000 item database.
1966: NASA awards parallel funding to Lockheed, for an in-house
installation, and to Bunker-Ramo, for a dial-up service, to automate
the NASA database.
1967: NASA issues an industry-wide RFP for NASA RECON; Lockheed
wins the contract, retains rights to Dialog software created.
1968: System Development Corporation (SDC) builds CIRC and COLEX
for Wright-Patterson Air Force Base; contracts with USOE for research
and dissemination of educational information (ERIC); does a "road
1969: SDC creates ELHILL, a retrieval program, for the National
Library of Medicine (NLM); tests software first with a Parkinson's
disease database, and then with 150,000 medical journal articles
1969: Lockheed installs systems at ESRO, for NASA RECON, and at
the Atomic Energy Commission, for Nuclear Science Abstracts; contracts
with USOE to provide leased-line service to ERIC at Stanford.
1970: NLM tests the teletypeinstead of leased linesfor
low-cost, low-speed information delivery.
1971: SDC creates ORBIT, written in PL/1; NLM runs ORBIT in-house
and expands AIM TWX into MEDLINE; an SDC computer backs up NLM's
installation of ORBIT.
1971: SDC loads ERIC database and offers SDC/ERIC publicly.
1971: Roger Summit and Dick Kollin negotiate online pricing for
Pandex database, at the ASIS national meeting.
1972: Threatened with loss of the NLM backup contract, Carlos
Cuadra surveys the commercial market for online services (survey
results indicated there was no market).
1972: Cuadra struggles to get CA Condensates database loaded onto
1972: At Lockheed, Roger Summit offers Dialog as a commercial
online service with NASA RECON, Nuclear Science Abstracts, ERIC,
and Pandex databases.
1973: January: Cuadra takes SDC Search Service "on the air."
Later Corporate History
1978: Carlos Cuadra leaves SDC and forms Cuadra Associates.
1981: Dialog Information Services, Inc. becomes a subsidiary of
1987: Robert Maxwell (Pergamon Press) purchases Orbit, renaming
it Pergamon Orbit Infoline.
1988: Dialog Information Services, Inc. acquired by Knight-Ridder,
1989: Robert Maxwell purchases Bibliographic Retrieval Services
(BRS); Orbit and BRS renamed Maxwell Online.
1991: Roger Summit retires from Dialog.
1994: French online host Questel purchases Orbit; names the composite
1995: Dialog becomes Knight-Ridder Information, Inc. (KRII)
1997: M.A.I.D. plc acquires Knight-Ridder Information, Inc.; forms
The Dialog Corporation.
2000: The Thomson Corporation purchases the Information Services
Division of The Dialog Corporation.
The Interview Begins
Today we want to concentrate on the industry. We're intrigued with the
development of what we are now calling the "first online" industry. Since
there are so many people who think the Internet started online, one of the
goals of this series is to show people that there was an online before the
Internet. We also want to appeal to the practitioners who lived through the
first online industry in various ways. So, we're looking at it from a business
point of view: How does an industry get developed? How did it come to be
Please describe how you each separately got into online. We know that
you worked within defense companies. Describe the funding and original purpose
of the projects from which online systems eventually evolved.
Cuadra: The point where I can remember noticing a beginning to our
being in the online business at SDC was when we did an experiment for the Federal
Technology Division at Wright-Patterson Air Force Base in Ohio. We had built
for them something called a Centralized Indexing and Reference Control
system (CIRC). It was mostly manual, kind of automated, and they liked it.
Then somehow we came to the idea of doing an experiment with a system that
was more interactive. That led to the second project, called COLEX the CIRC OnLine EXperiment.
All this was government-funded. Then, about the same time, in 1968, we got
a contract with the Department of Education to conduct some research on their
ERIC centers, to write a book, and to prepare brochures on dissemination and
use of information.
ERIC and MEDLINE
Cuadra: As part of that contract or maybe a follow-on, we were asked
to do a road show to demonstrate online retrieval to the educational community.
Two people from my staff did that. They carried around a teletype teletype
machines were so big then, they had to be shipped in two boxes! Fortunately,
one of the staff members was large and strong. They took this around to different
cities and demonstrated online searching at something between six and 10 characters
a second. Clunk, clunk, clunk, clunk, clunk.
About the same time, we were asked by the National Library of Medicine to
test some new software we had developed, using a Parkinson's disease database.
We demonstrated the database, and in May 1970, that project led to the start
of a service called AIM TWX that we ran for NLM. The name stood for Abridged Index Medicus, which
is a publication, and Teletype Writer eXchange, which
is a service. The idea was to provide service to medical libraries, teaching
hospitals, and other health professionals using a database of 130,000 records
that were a subset of the MEDLARS database.
The ORBIT Software
Cuadra: In 1971, we completed a version of new software called ORBIT.
It was SDC-funded and was written in PL/1, which our programmers tell me was
regarded as an exotic new language at the time. We had a lot of fuss with SDC
management to get them to let us use it. NLM licensed a version of this new
software, which they referred to as ELHILL (to honor Senator Lister Hill, a
major NLM supporter), and used it to start a service called MEDLINE (MEDLARS
On Line) that provided access to the full range of documents in the MEDLARS
database. A few months later, they began providing MEDLINE service on SDC's
computer as a backup. NLM allowed SDC to provide access to this backup MEDLINE
service to members of the Pharmaceutical Manufacturers Association (PMA) because
the PMA companies were doing some indexing for NLM, and as the quid pro quo,
they got this access. So commercial companies as well as health professionals
were now searching MEDLINE.
The response to MEDLINE was very good. When there was too much load, people
could get to MEDLINE on SDC's computer. I think the PMA companies at that time
used only SDC's MEDLINE.
Cuadra: In late 1971, even before we started our MEDLINE service for
NLM, some of us got the idea of selling other databases, too. The one we picked
was ERIC, because we knew so much about it from the work we'd done for the
U.S. Office of Education. I'm not sure whether we bought it or got permission.
Summit: It was free.
Cuadra: It was free. So we loaded that database and started selling
access to it under the name SDC/ERIC, and this was the beginning of our commercial
Summit: When was that? 1971, do you think?
Cuadra: Late 1971. There are two reasons why I think it was 1971.
One of our programmers found a brochure, the back of which shows a little code
that says '71. So the brochure was done in 1971, though I can't remember when
we actually sent it out into the world. Also Ted Brandhorst wrote a chapter
in the Annual Review of Information Science and Technology (volume 7,
1972), where he said, "SDC inaugurated in late 1971 an online commercial search
service, SDC/ERIC." He followed that with, "At this writing," which was in
1972, when he was reviewing the literature of 1971, "DIALOG is implementing
plans to 1) offer the public online retrieval services against the ERIC database
on a commercial basis and 2) add natural language text searching."
The First Market Research
Cuadra: Sometime in 1972, I believe, NLM notified SDC that it no longer
needed to use the SDC computer as the backup for MEDLINE. So, the funding for
that service was going to stop. During that year, I think, I surveyed about
7,000 users of information from NTIS. I got the list from NTIS, and sent out
a very simple survey form two pages trying to explain what "online" meant,
because most people weren't using the term then, and asking questions such
as, "If this thing were to exist, and you could use it, would you find it useful?" And
I may have asked what they would pay for it.
Summit: You didn't.
Cuadra: OK, I guess I didn't. I got about a 1 percent response 70
or 72 returns. Most of them said, "I don't know" or, "Ho-hum." A few said, "That
might be nice." In effect, considering the 7,000 survey returns and looking
at the results, it said, "There is no business here." I looked at the data
for a while and thought, "Maybe my survey's dumb." I was a Ph.D. psychologist
and I thought I knew how to do surveys. But I finally decided that my survey
was wrong, or the respondents were wrong. I decided, "Once we do it, they'll
want it." So I put the survey away locked it in a drawer and
went to management and said that we had to start an online retrieval service.
We started almost by accident.
Summit: Just to comment on that survey, because it was a very important
thing. We at Lockheed had been doing all the NASA RECON system development;
we had the Atomic Energy Commission (AEC) contract, and we had ERIC up and
operating under a contract from the U.S. Office of Education for the ERIC facility.
We got the thought, very quietly, that by golly, maybe there's a commercial
application here. One of the defense contractors asked us if they could search
AEC/RECON, or AEC, which was known as Nuclear Science Abstracts (NSA) at that
time. We struggled with the Lockheed folks on the issue of having a nongovernment
entity accessing and searching the database that had to be about 1970
or 1971. We did a contract with them I think it was General Dynamics
in San Jose and they were really our first nongovernment customer.
The Survey Comes to Lockheed
Summit: But what happened then was this: Lockheed of course was an
NTIS user, so our library received a survey, and our librarian brought it in
and said, "Hey, look at this." I looked at it and it was a survey from one
Carlos Cuadra, whom we didn't know and had never heard of, at SDC, and I thought, "Oh,
shit. Somebody else has this idea." We clearly had the idea of commercializing,
and we felt we were going to leapfrog over the other SDI services that were
around, spinning a lot of tape. So that survey really stimulated me into action
with our management. I said to them, "We have to get this...."
Cuadra: You used this as an excuse!
Summit: You bet! I said "We have to get this thing going. We have
an edge now, because there are no services out there. But here this guy down
at SDC is going to start a commercial service...."
Was that really the first time you heard of Carlos Cuadra?
Cuadra: Well, I knew of Roger's name because some of our people had
gone to a conference on the man/machine interface, or the machine/user interface,
and Roger was there; I think he was a speaker.
Summit: Well, I'm sure I did know Carlos.
Cuadra: I was the editor of the Annual Review at that time.
Summit: So, I knew Carlos, but I didn't know what he was up to. But
at that point, I could go to management and say that this guy down there has
proved there might be a business, so we have to get going. They said, "Fine.
Go for it."
Dr. Cuadra, do you have the original survey?
Cuadra: Nope. I don't have any paper from SDC. We were not a very
visible part of the company, so I doubt that it's in an archive. In fact, I
wasn't even running SDC's MEDLINE service then. That was done by other staff
A Go-Ahead from Management
Cuadra: To finish the sequence of SDC getting into commercial operations....
I went to management and said that we ought to start an online service. They
were aware that the funding from the NLM contracts for us to run this computer
as a backup was going to go away. The only other database we were offering
at that time was the ERIC database. We were kind of piggybacking. Since the
software was available, all we needed to do was add other data that were relevant
to other people. So with no NLM funding, we had to either fish or cut bait forget
the whole thing or move ahead and create something. I argued for moving ahead
and creating something, and I was given an OK to start it up to see what we
could do in 3 months.
Just 3 months???!!!
Cuadra: Yes, and you remember how long it took to get people online
and doing things. Management also tied my hands by saying they wanted to make
sure that people didn't just fool around with our computer; they wanted to
get at least $500 a month, so customers would have to commit in advance. I
said customers wouldn't do that, and management said, "They'll have to." SDC
worked for the Air Force and they were used to issuing contracts for a million
dollars, $50 million, $80 million big invoices for the work of hundreds
of programmers. The idea of picking up $23.89 for a search was totally foreign.
So they wanted to make sure they got money up front.
It took 3 months before I convinced them that their rule was getting in the
way of starting up. The librarians we went to had no idea of what they were
going to do, had no idea how much they were going to spend, didn't want to
have a terrible surprise, and just wanted to get started and try the service
out. So eventually that $500 number dropped out of the picture.
So the flat-rate price that was indicated at that point may have worked
if it was tiny, but it was way too high?
Cuadra: Yes, it was way too high for people who were just getting
Summit: Was that price ever publicized? Did you try to start the service
with that $500 charge, or did you get it turned off before you started?
Cuadra: We started with it.
Summit: This ties into one of Dialog's philosophies, and you may have
been the trigger for that. What I said was, "No minimums, no monthly fees."
The First ORBIT Databases
Cuadra: Yes, you have to get them hooked first. When we first went
on the air, we had to decide what databases to provide. We couldn't use the
NLM MEDLINE database, because we didn't have the rights to it. We were using
a copy that belonged to PMA as their quid pro quo for their indexing work,
and I think NLM wouldn't agree to leave it on our computer to serve a commercial
market. So that disappeared, and we had only this piddling little ERIC database.
That was it!
Bjørner: That's the reason I originally went
online, I will tell you parenthetically ERIC.
Summit: On SDC or Dialog?
Bjørner: BRS. [Laughter]
Summit: That was later.
Cuadra: Interview's over, right?
Converting CA Condensates
Cuadra: So, the fun part of this, I think, is how we survived the
first acid test. In late 1972, when we got the OK, a journeyman programmer
was assigned to write a program to convert the data. We chose, as our first
major commercial database, Chemical Abstracts Condensates. The programmer worked
on it for a little over a month, and his program converted about 4,000 records
an hour. I did the computations I think we had a half a million records
to go and realized that we had no way to do the conversion. We were
out of business before we even started because our computer center charged
us commercial rates!
What we're talking about here is the program that would convert the data
from Chemical Abstracts Condensates into ORBIT's database. It ran on a mainframe
and converted record, record, record, record, record. It was primarily header
information, some brief indexing, and I don't recall whether or not it had
abstracts at the time.
Summit: No. Well, we were never able to get abstracts from the devils.
Cuadra: Yes, I remember. I went home as depressed as possible. I sat
there at the dinner table with my head in my hands, with my two sons and my
wife, talking about this disaster; the thing that I wanted to start was doomed.
And one son, who was 19 at the time, said, "Well, maybe I could do something." He
had just learned to program, had been at it for maybe a year, and I said, "If
you can do that, I'll reward you." I built a matrix at the dining room table,
and it was based on the number of days to write the program and the rate of
database loading. If he took one day and his program was able to load 50,000
records an hour, he would get a Mercedes Benz. At the other end, if he took
10 days and it loaded only 6,000 records an hour, he would get a Matchbox car.
And it turned out somewhere in the middle. He wrote a program that converted
20,000 records an hour; he did it in about 6 days; and I had to buy him a Toyota.
He saved SDC's search service. If he had not done that, it never would have
started. So we got past that day. I paid him myself; SDC didn't.
Who's Who: Key People
Brandhorst, Wesley Theodore (Ted) Assistant director
of the NASA Scientific and Technical Information Facility from
1962 until 1969. In 1970, became the director of the ERIC Processing
and Reference Facility, the centralized database processor for the
U.S. Office of Education's Educational Resources Information Center
Burchinal, Lee Deputy head, Division of Educational
Research, U.S. Office of Education, 1965. Established the ERIC Clearinghouses
Day, Melvin Sherman 1946: Joined the Atomic Energy
Commission (AEC) at Oak Ridge. 1958: Director of the Technical
Information Division of AEC. 1960: Deputy director, Office of Technical
Information and Education, NASA. 1962: First director, Technical
Information Division. 1971: Head, Office of Science Information Service,
National Science Foundation (NSF). 1972: Deputy director, National
Library of Medicine (NLM).
Luhn, Hans Peter Worked for IBM from 1941 until his
retirement in 1961. In 1958, Luhn's interest in applying machines to
literature data processing led to his development of Key Word in Context
(KWIC) indexing and selective dissemination of information (SDI).
He died in 1964.
Marron, Harvey Chief, Educational Research Information
Center (ERIC), in the late 1960s and early 1970s.
Cuadra: So that's kind of how we started and then we, like Dialog,
later began adding databases. We'd ask each group of customers, for example the
pharmaceutical folks "What do you need next? What would make this service
more useful?" Then we'd go out and track it down. But that's the start; that
was in late 1972, and I think we went on the air in January 1973, plus or minus
a few days.
Now we want to see what was happening in parallel over at Dialog.
Summit: I went to work as a summer hire for Lockheed in 1960. I was
doing my Ph.D., between the orals and the dissertation, and I needed a summer
job. I was assigned two project areas. One was computer simulation and the
other was information retrieval. I worked on those for about 4 years, until
1964. I was working mainly in simulation, because that's where my dissertation
was, and taking some management classes.
Then in 1963 or 1964, we were doing some small information retrieval experiments.
I wrote a program called MATICO, which was Machine-Aided Technical Information Center.
I thought it was cataloging, but really it was a little program so that you
could put in the data once and print all the catalog cards that you needed
for the card catalog. So, I started getting involved in information retrieval
at that time.
Retrieval was very crude in those days; it was all done by batch processing,
tape spinning. You'd have a bunch of retrieval queries on punched cards and
your database on tape, and you'd start processing that tape one record at a
time and matching against the queries that were on cards.
Early "Alert" Services
Summit: Let me go back a little bit. That was the way retrospective
retrieval was done, but more important were SDI services, Selective Dissemination
of Information, as we called it then. Later it got called "alert...."
Cuadra: Now it's "push technology"!
Summit: Right. SDI was invented, or developed, by Hans Peter Luhn,
in 1960 or so. At least he formulated the idea. There were several universities
that were offering services internally on certain of the tapes. Luhn was at
IBM, and he was doing SDIs within IBM; I think they were on their internal
literature and maybe some journal articles. We talked our management into setting
up the Information Sciences Laboratory at the time the IBM 360 was announced
(1964). The 360 was so-called third-generation technology. What I was describing
before was second-generation technology, the tape-spinning stuff.
We looked around at the batch retrieval activities of the day. We couldn't
get in, we couldn't get a paper accepted, we couldn't seem to get admission
into the information retrieval clique at the time. I'm talking about meetings
such as ASIS or the American Documentation Institute, as it was called then.
I remember feeling very frustrated about that. They weren't interested; they
had an information focus, not a computer focus. I can't say we tried real hard,
but I remember having the feeling that we were outsiders.
WHAT'S WHAT: Names, Acronyms,
AEC Atomic Energy Commission. In operation from 1947
until 1975. Replaced by the Energy Research and Development Administration
(ERDA). In 1977, ERDA was integrated with the Federal Energy Administration
and other federal energy functions to create the U.S. Department of
AEC/RECON REmote CONsole: Atomic Energy Commission's
online bibliographic system, used to create the Nuclear Science Abstracts
AIM TWX Abridged Index Medicus; Teletype Writer
eXchange. Online bibliographic system developed for the National Library
of Medicine by SDC.
ASIS American Society for Information Science. Founded
in 1937 as the American Documentation Institute (ADI). Name changed
to ASIS in 1968 and to the American Society for Information Science
and Technology (ASIST) in 2000.
CIRC Centralized Indexing and Reference Control. Online
bibliographic searching system, developed by SDC in 1966 for the Federal
Technology Division at Wright-Patterson Air Force Base.
COLEX CIRC OnLine EXperiment. Second version of CIRC
developed by SDC.
Dialog 1972: At Lockheed, Roger Summit offered DIALOG
as a commercial online service. 1981: Dialog Information Services,
Inc. became a subsidiary of Lockheed Corporation. 1988: Dialog was
acquired by Knight-Ridder, Inc. 1995: Dialog became Knight-Ridder Information,
Inc. (KRII). 1997: M.A.I.D. plc acquired Knight-Ridder, merging to
form the Dialog Corporation. 2000: The Thomson Corporation purchased
the Information Services Division of the Dialog Corporation.
ERIC The Educational Resources Information Center (ERIC).
National information system providing access to education-related literature.
Established in 1966, ERIC is supported by the U.S. Department of Education's
Office of Educational Research and Improvement and is now administered
by the National Library of Education (NLE).
Information Access Corporation (IAC) Company founded
by Dick Kollin in 1976. 1980: Ziff-Davis acquired the company. 1994:
IAC and Predicasts consolidated under the IAC name. 1995: Thomson Corporation
bought IAC. 1998: Gale Research, IAC, and Primary Source Media merged
to form the Gale Group under the Thomson umbrella.
Lockheed The Lockheed Corporation was formed in 1932.
Lockheed Corporation and Martin Marietta Corporation merged in 1995
to form the Lockheed Martin Corporation. In 1968, NASA contracted with
the Lockheed Missile and Space Company to manage large data files.
The following year, Lockheed's Information Sciences Laboratory demonstrated
an interactive retrieval service. See Dialog.
MEDLARS Medical Literature Analysis
and Retrieval System. Computerized bibliographic system,
originally used in the National Library of Medicine (NLM) and named
by NLM's Frank Rogers and Seymour Taine in 1960. MEDLARS was designed
by General Electric, which completed the system in 1964. MEDLARS II
was designed and developed by SDC, which completed the system in 1974.
MEDLINE MEDLARS onLINE. Online system
of indexed journal citations and abstracts developed for users outside
the National Library of Medicine (NLM) in 1971. MEDLINE is the major
component of NLM's PubMed database, now searchable via the Internet.
NASA National Aeronautics and Space Administration.
Established in 1958, from the earlier National Advisory Committee for
Aeronautics (including three major research laboratories Langley
Aeronautical Laboratory, Ames Aeronautical Laboratory, and Lewis Flight
Propulsion Laboratory and two smaller test facilities).
NASA/RECON REmote CONsole. NASA's "first multi-site" online
bibliographic system, created in 1968.
NLM National Library of Medicine. Organized under the
U.S. Department of Health, National Institutes of Health (NIH). For
more than 100 years, the Library has published the print Index Medicus,
a guide to journal articles. See also MEDLARS and MEDLINE.
NSA Nuclear Science Abstracts. Collection of international
nuclear science and technology literature for the period 1948 through
1976, including scientific and technical reports of the U.S. Atomic
Energy Commission (AEC) and the U.S. Energy Research and Development
Administration (ERDA) and its contractors, other agencies, universities,
and industrial and research organizations. NSA is the precursor of
the Energy Science and Technology Database.
ORBIT In 1969, the System Development Corporation (SDC)
created the ELHILL retrieval program for the National Library of Medicine
(NLM). ORBIT, a commercial offshoot of ELHILL, became publicly available
in 1972. Robert Maxwell (Pergamon Press) bought ORBIT in 1987 and renamed
it Pergamon Orbit Infoline. In 1989, with Maxwell's purchase of Bibliographic
Retrieval Service (BRS), the entire group was renamed Maxwell Online.
In 1994, Questel, the French-based online host, bought Orbit and named
the composite company Questel-Orbit.
Pandex In 1971, Pandex was the first commercial, nongovernmental
database made publicly available by an online search service (Dialog
PL/1 Programming Language 1. Third-generation programming
language developed in the early 1960s as an alternative to assembler
language (for low-level computer processing functions). PL/I was an
antecedent of the C programming language, which basically replaced
it as an all-purpose programming language.
PMA Pharmaceutical Manufacturers Association. Former
name of the Pharmaceutical Research and Manufacturers
of America (PhRMA). Founded 1958.
SDC System Development Corporation. Evolved out of
the Systems Development Division of the RAND Corporation. The Division
was spun off in 1957 and became the nonprofit company System Development
Corporation. In 1968, SDC became a for-profit operation, acquired by
Burroughs Corporation in 1980.
SDI Selective Dissemination of Information. Automatic
method to provide current awareness services. Brainchild of Hans Peter
Luhn, who introduced this "business intelligence system" in 1958.
Teletype An early data and record communications interface
system. Developed by the Morkrum Company in 1906. 1929: Morkrum changed
its name to the Teletype Corporation. 1930: Teletype was purchased
by the Bell System and became a subsidiary of Western Electric. 1984:
the divestiture of the Bell System resulted in the replacement of the
Teletype name and logo with the AT&T name and logo.
TWX Teletype Writer eXchange. Early data network, started
by AT&T in 1931. The teletype switching service placed calls automatically
to other teletype machines. The first computer time-sharing was conducted
in the mid-1960s over the TWX network. AT&T sold TWX to Western
Union in 1972.
Tymshare, Inc. Company that created the commercial
TYMNET computer network in 1971.
U.S. Office of Education Created in 1868, operating
under different titles and housed in various government agencies, including
the U.S. Department of Health, Education, and Welfare (now the U.S.
Department of Health and Human Services). In 1980, the current U.S.
Department of Education was formed.
Summit: My feeling about the third-generation computer technology
was we can leapfrog this whole business and provide access in a way
it hasn't been provided before. So, we got Lockheed to buy an IBM 360/30 computer.
We got the third such computer that IBM manufactured. We set it up, and I was
in charge of the information retrieval experiments. The Information Sciences
Laboratory was to examine the different opportunities that existed with third-generation
hardware in the area of information.
Cuadra: Was all this Lockheed-funded?
Summit: It was all independent research-funded. The largest database
I knew about was the NASA database, which was 300,000 citations no abstracts just
citations and indexing. Mel Day was well known for his work in developing the
NASA database. He has a history of contact with most all the government information
activities of the time. He later was at the National Library of Medicine and
then at NSF (the National Science Foundation). He really made the rounds.
The Road to NASA Funding
Summit: I went to see Mel Day in Washington and said, "We have this
new technology that's really great." And he said, "Look, Roger, I have a couple
of dozen people a week come in here and tell me their computer programs can
do everything but read my mind in terms of information retrieval. You have
to demonstrate something to me."
So, I went back to Lockheed. I wasn't discouraged, because we had a prototype
operating then on a small database. I wrote an unsolicited proposal to NASA
to do a pilot demonstration of online information retrieval using the NASA
database. That had to be about 1965 or 1966. The reason I did this is because
there were a lot of what we called little toy databases around that people
were doing things with. But, if retrieval was going to mean anything, it had
to be done on a very large database. I figured, if we could demonstrate on
a 300,000-citation database, then the technology we were using would be extendable.
Mel went for that. When we submitted our unsolicited proposal, Mel said it
was fine, but then he said, "Oh, by the way, Bunker-Ramo's going to do a parallel
project." So Bunker-Ramo also received an award from NASA to do a prototype.
We installed a terminal in NASA headquarters, but NASA specified a dial-up
teletype protocol for Bunker-Ramo. After 2 years, the Bunker-Ramo experiment
proved unfruitful and NASA dropped it. So, the Bunker-Ramo project died, and
Then in 1967, NASA came out with an industry-wide RFP for the NASA/RECON
System and specified 20 system requirements. Dialog included, at that time,
19 of those requirements. Consequently, we were able to bid quite reasonably
on that contract. I remember going to the proposal briefing conference. All
the big companies were there ... the Computer Sciences Corporation and others.
We bid against probably six or eight companies, and then won the development
contract for NASA/RECON. We wrote a Rights and Data clause in that contract,
which said that we would hold the rights to use the software we developed.
NASA started up with their database on the system that we installed.
Then, also in the 1967 time frame, we negotiated a contract with ESRO, the
European Space Research Organization, to install a similar system in Europe
with the NASA database, which we did, probably in 1968 or 1969. We called it
Dialog, but they called it NASA/RECON. They had to do that for political reasons,
to get the NASA database. We also installed a system at the Atomic Energy Commission,
using the Nuclear Science Abstracts database. They were running the Dialog
retrieval system in their facility, on that database.
ERIC and the Services Business
Summit: I read about the ERIC database, and thought, "This is a natural." They
were doing some investigations on whether they could automate access to the
ERIC database. So, I went back and talked with Harvey Marron and Lee Burchinal
at the U.S. Office of Education. I said, "Look, you don't have to fool around
with this yourselves; we have a system that will run your database." I wasn't
aware, not consciously aware, of just where SDC stood on the ERIC database.
Cuadra: This was around the same time that we did our road show (1969)
demonstrating the ERIC database.
Summit: I think we had it online at that time. We were going to do
the same thing that we did with AEC and with ESRO and with NASA. I said, "We'll
install the system on your computer and then you can run the ERIC database." And
they said, "We don't want to mess around with computers. You put it on your
computer, and then we'll pay you to operate it on your computer, to access
it from our clearinghouses around the country." They preferred to have us run
the computers. DIALOG did leased lines then. Our first installation was at
Stanford, which was kind of easy. That must have been in 1968 or 1969. A fellow
named Don Combs was in charge of the Stanford center.
And then ERIC said, "Well, we want to expand the access centers. We want
a terminal headquarters in Washington and also up in Pennsylvania." ERIC was
really a milestone for us because it moved us out of the systems development
and installation contracting kind of business into the services businesses,
and we saw how sweet that was. With a contracting business, you die a thousand
deaths. Each time a contract term concludes, you have to go out and get reborn
again. But with the services business, once you get folks hooked on it, there's
continued, recurring revenue.
Pandex and Pricing
Summit: We had three databases then...NASA , Nuclear Science Abstracts,
and ERIC. And then I met Dick Kollin at that famous 1971 ASIS meeting; he had
something called the Pandex database.
Kollin: Pandex was a general science database of 1,200 serials, very
similar to SciSearch.
Summit: I remember Dick telling stories of using garment carts to
wheel the big tape reels up and down the streets of midtown Manhattan. I forget
why they had to be transported, but it gives you an idea of how humongous they
Is that where you two met? Or just where you two came up with online
Summit: It could have been both, because I knew of the Pandex database
before that meeting.
Kollin: We were sitting on a set of bleachers...
Summit: ...a stairway...
Kollin: ...long into the night, talking about doing it by the drop
or by the bucket.
Summit: That's right! I wish I could remember the details of that
negotiation. I think Dick wanted to get paid by the item, and I wanted to get
paid by elapsed time, because that was the way we billed our service. It was
kind of a consignment thing: We'll take the database on and we'll pay you a
percentage or so much per hour that we collect from our customers. There's
no risk associated with that. With Lockheed, I had to be very risk-adverse,
as opposed to going to them and saying, "I want $50,000 to lease this database
for a year," which I never could have gotten. So that kind of set the stage
for online pricing; we didn't charge for citations until years after that.
Kollin: I set it up by wall time.
Summit: Wall time. Clock time. So that was your idea, Dick?
Summit: Bad news. Okay, I can blame Dick now.
So you adopted this pricing plan for the Pandex database when you put
it onto Dialog?
Summit: Yes, and we kind of take the year of 1972 as the birth date
of all of that; it was largely because Tymshare had set up a telecommunications
network and we could get away from the leased lines. We could have a much broader
service than with our leased line service, which was continuing by contract.
With leased lines, we didn't charge by the hour; we just had a contract of
so much per month.
The Teletype Network
Cuadra: There is another point to make about AIM TWX. In May 1970,
NLM decided not to follow the model that Roger and others were using of the
dedicated leased lines and fancy terminals. They were thinking of providing
information to libraries and teaching hospitals and health professions, who
didn't have much money to buy information. NLM wanted to test the acceptance
of low-cost, low-speed information delivery. The teletype became their way
of testing the waters.
Summit: Yes. Teletype was a network ... way back when, over the railroad
lines ... it was really the only network that was around. The point I was going
to make there was, again, risk-aversion. We had all these contracts,
NASA and so on, and an ERIC contract that paid the bills. We were able to go
into commercial business on an incremental basis because we had all our costs
covered in our fixed contracts with these government agencies. So we thought, "Well,
let's try it."
Cuadra: Does that mean you could use the computer they were paying
for to generate income?
Cuadra: That's what we couldn't do.
Summit: But they weren't paying for the computer; they had a contract
that said, "Lockheed Dialog will load and maintain your database...."
Cuadra: They were subsidizing you, in effect. You had the computer
and you were using it.
Summit: Well, no, they weren't subsidizing that's a bad word.
We had a contract with them, and we had a price on that contract. Yes, we paid
our computer costs out of that price, but it wasn't that they were paying computer
charges per se. It was a fixed price contract.
Cuadra: It didn't prevent you from using the very same box to do other
Summit: As was your case. We had a real advantage at the gate, price-wise
and cost-wise, because this computer helped. This computer was "independent
research" and it was still owned by Lockheed. So the revenue we were getting
was just contract revenue coming into Lockheed.
It really was a different management philosophy.
Summit: Totally different. It was just circumstances. That was the
way it happened to be set up. The key was that we got Lockheed to buy this
computer on independent research for the Information Sciences Lab, and we were
simply generating revenue from this Lockheed-owned computer.
The Corporate Culture
Now we're at the stage where there are two systems out there (Dialog and
ORBIT) beginning to go public and hoping for revenue and more customers.
You're still talking about big organizations that are involved and could
afford the costs
Cuadra: Elements of big organizations.
Summit: Very small elements of very large organizations. We were just
kind of set aside. We were in a closet somewhere, and nobody paid much attention
to us. We didn't have to make a profit, but if we ran some revenue in, they
were happy. I knew that we had to get off independent research that's
why I wanted to go out and get these contracts early on. Lockheed didn't know
what information retrieval was. They were used to banging metal and things
like that. I knew if I could get government contracts they knew what
government contracts were that would rationalize the business.
Cuadra: SDC understood information retrieval in the sense of radar
returns, plots of flights crossing the country, handing over data from one
air defense sector to another they didn't understand retrieval in terms
of text. Documents, serials, publications.
It is very interesting that these systems that contained real information,
not data text and so on came out of that environment. It seems
in spite of them, rather than because of them.
Cuadra: Yes, and SDC Search Service wasn't the first real text-retrieval
system. For at least 6 or 7 years before that, there were various researchers
who were playing with things. One wrote a program called Proto-synthex that
provided access to the Golden Book encyclopedia. Years later, the corporate
secretary, in a fit of pique or in a fit of cleanup, threw out 50,000 cards,
which were the only source code to this program. I still cringe.
Summit: We had some corporate support. There was an executive vice
president, Herschel Brown, who read the report on the automation of the Library
of Congress, which we refer to as the Red Book. He said, "Ah, automating the
Library of Congress is the kind of task that a company like Lockheed could
take on." So he backed us in the Information Sciences Lab, and it wasn't a
total out-sell. We had support from him until he died. He was very important.
I won't speak for Carlos, but I think this is true for both it isn't
that Lockheed was interested in information retrieval or that System Development
Corporation was interested in information retrieval; it was that Carlos Cuadra
was interested in information retrieval and Roger Summit was interested in
information retrieval. We each saw some future to it, and then we had to figure
out how to encapsulate that business within these vast bureaucracies.
issue, the conversation among these online pioneers continues, addressing
issues of competitors and colleagues, the database race, marketing and
building the industry, and business vs. government.
"About ERIC," http://www.eric.ed.gov/about/about.html.
American Society for Information Science & Technology (ASIST), "ASIST
and Its Members" [http://www.asis.org/AboutASIS/the-society.html]; "Day,
Melvin Sherman," [http://www.asis.org/Features/Pioneers/day.htm]; "Luhn,
Hans Peter" [http://www.asis.org/Features/Pioneers/luhn.htm];
and "National Aeronautics and Space Administration"
Basch, Reva, "Twenty Years of Headline News," ONLINE, vol.
21, no. 1, January 1997, http://www.infotoday.com/online/JanOL97/headlines.html.
Bourne, Charles P., "Forty Years of Database Distribution and
Use: An Overview and Observation,"
1999 Miles Conrad Memorial Lecture, February 23, 1999, http://www.nfais.org/MilesConrad.asp?WhichYear=1999
Burchinal, Lee G., "The Tale of Two ERICS: Factors Influencing
the Development of the First ERIC and Its Transformation into a National
System," Journal of the American Society for Information Science,
vol. 51, no. 6, April 2000, pp. 567-575.
Charles Babbage Institute, University of Minnesota, Minneapolis, "Burroughs
Corporation Records (CBI 90): System Development Corporation Records,
Christiani, Linnea, "History Lesson: Magazine Index & Magazine
Database," ONLINE, vol. 27, no. 1, January/February 2003,
"Fact Sheet: The National Library of Medicine," January 30, 2003, http://www.libsci.sc.edu/bob/ISP/luhn.
"Fact Sheet: What's the Difference Between MEDLINE and PubMed?," July
10, 2002, http://www.nlm.nih.gov/pubs/factsheets/dif_med_pub.html.
Garber, Stephen J. and Roger D. Launius, "A Brief History of the
National Aeronautics and Space Administration," http://www.hq.nasa.gov/office/
"Hans Peter Luhn," http://www.libsci.sc.edu/bob/ISP/luhn.htm
"The History of Dialog," http://www.dialog.com/about/history/.
"IEEE History Center," http://www.ieee.org/organizations/
"Key Dates for Dialog," http://www.dialog.com/about/keydates/.
Lambert, Nancy, "Orbit and Questel-Orbit: Farewell and Hail," Searcher,
vol. 8, no. 2, February 2000, http://www.infotoday.com/searcher/feb00/lambert.htm.
"'Magazine Index,' a New Database Aimed at the Markets of the
Reader's Guide and the New York Times Information Bank, Due Online
in January ," ONLINE, vol. 2, no. 2, January 1978, pp.
Massachusetts Institute of Technology, Laboratory for Computer
Science, "TELECOM Digest and Archives," 1991, http://massis.lcs.mit.edu/archives/history/teletype.
Massachusetts Institute of Technology, Laboratory for Computer
Science, "TELECOM Digest and Archives," 1997, http://massis.lcs.mit.edu/archives/history/twx.1961.automation.
Miles, Wyndham D., A History of the National Library of Medicine:
The Nation's Treasury of Medical Knowledge, Washington, D.C.:
U.S. Department of Health and Human Services, Public Health Service,
National Institutes of Health, National Library of Medicine, 1982,
pp. 368-370, 384-385, 388-390, 422-424.
Mitchell, William. "The Genesis of NASA RECON," http://www.ualr.edu/~wmmitchell/NASARECON.htm and http://www.ualr.edu/~wmmitchell/
Nuclear Science Abstracts Database, http://grc.ntis.gov/nsa.htm.
Pemberton, Jeffery K., "An Interview with Dick Kollin of EasyNet," ONLINE,
vol. 10, no. 8, May 1986, pp. 17-24.
Provenzano, Dominic. "Where Are They Now?," ONLINE, (Special
10th Anniversary Issue), vol. 11, no. 1, January 1987, pp. 25-42.
"Some History," http://www.vauxelectronics.com/
Soy, Susan K. "Class Lecture Notes: H. P. Luhn and Automatic Indexing References
to the Early Years of Automatic Indexing and Information Retrieval," Spring
Summit, Roger. "Reflections on the Beginnings of Dialog: The Birth
of Online Information Access," Chronolog, June 2000, http://support.dialog.com/publications/
U.S. Department of Energy, Office of Science, "History," http://www.ch.doe.gov/insidech/history/index.html.
"What Is the U.S. Department of Education?," http://www.ed.gov/pubs/overview/whatis.html.
WhatIs.com, http://whatis.techtarget.com/ (dictionary
of information technology words).
Susanne Bjørner is an independent consultant to publishers and
authors and writes about the information professions and industry. Contact her
Stephanie C. Ardito is the principal of
Ardito Information & Research, Inc., a full-service
information firm based in Wilmington, Delaware.
Contact her at email@example.com.