Elizabeth Feinler and The History of the Internet
“I am proud to have had a small role in the development of the Internet, a technical phenomenon that has changed the way the world learns and communicates.”
Elizabeth “Jake” Feinler is a pioneering information scientist and the former Director of the Network Information Systems Center at the Stanford Research Institute (now SRI International), which she joined in 1960 to work in the Information Research Department.
In 1972, Elizabeth joined Doug Engelbart’s Augmentation Research Center (ARC) to work on the ARPANET Resource Handbook. Elizabeth was principal investigator for the Network Information Center (NIC) project from 1974 to 1989. During that time, Elizabeth’s NIC group worked on the Advanced Research Projects Agency Network (ARPANET), which evolved into the Defense Data Network (DDN), both forerunners to the Internet .
The NIC provided ARPANET users with various support services, a directory, a resource handbook (list of services), and the DoD protocol handbook. It was also Elizabeth’s group who managed the first host-naming registry for the Internet and developed the top-level domain naming system of .com, .gov, .org, .edu, and .mil, which is still in use today.
Elizabeth became the director of the Network Information Systems Center at SRI in 1986. She left there in 1989 to work for Sterling Software Corp.
At Sterling Software Corp., Elizabeth worked as a contract network requirements manager at NASA Ames Research Center, where she helped develop guidelines for managing the NASA Science Internet (NSI), NIC and the NASA websites.
Elizabeth was inducted into the Internet Hall of Fame in 2012. Since retiring, she has been a consultant at The Computer History Museum, where she donated, organized, and detailed over 350 boxes of archives from the Engelbart and NIC projects.
Elizabeth was the first in her family to graduate from college, which she did in 1954 with a Bachelor of Science degree in chemistry from West Liberty University.
This interview is an extract from my 2016 interview with Elizabeth for my book Female Innovators at Work.
Danielle Newnham: Can you tell me about your background? What were you like growing up?
Elizabeth Feinler: I grew up in Wheeling, West Virginia, on the island during the time of The Great Depression. Money was very tight and no one in my family was wealthy. Both of my grandfathers and my step-father’s father lost their businesses in the Depression. My father became an alcoholic, so my parents separated when I was about two years old. We then went to live with my mother’s parents who ran a boarding house, and my mother then worked as a clerk in a shoe store. When I was seven, my mother remarried. My sister, half-brother, and I lived with my mother and stepfather, a steel worker, until I went off to college at eighteen.
My life was a little schizophrenic as, early on, I spent weekends with my father and paternal grandparents and week days with my mother and her parents. Life at home was organized. I had chores, was expected to help out, be responsible, be courteous to the elderly boarders, and do well in school. No helicopter mom here. If I got a bad teacher or thought I got a bad deal at school, the message from home was, “Deal with it. No one’s life is without problems.”
On the other hand when I was at my father’s house, all stops were out. Movies, ice cream and candy, circuses, favourite foods, presents — you name it. My sister and I were spoiled rotten there. My grandfather had owned several theatres in town, and even though he had lost them in the depression, and had to go back to work in one as a ticket taker to make ends meet, as a courtesy to him we were allowed to go to all the movies in town for free. I was interested in art and my uncle, who was a commercial artist, gave me paints, brushes, and paper he no longer needed. Dad made us a doll house, desks with swivel chairs, and games for us to play, and my two aunts gave us old clothes and costume jewellery to play dress up or “movie star.”
Needless to say I had a rich Walter Mitty–like fantasy life. The movies had some strong women in them then — Betty Davis, Ingrid Bergman, Mae West, Joan Crawford. Barbara Stanwyck, Carole Lombard, Vivien Leigh — and they served as role models because I saw so many movies — sometimes eight or ten in a weekend.
I went to public schools, liked school, and did well. I was shy, but had many friends. I met one really good, lifelong friend in second grade and am still good friends with her today, although she now lives half way across the country. We went to grade school and high school together, were college roommates, I was in her wedding, and we worked our way through college together as assistant librarians in the college library.
As kids, we studied together and played together. We liked nature study and went on nature hikes and roamed the river bank or went to the library. Kids weren’t allowed to read adult library books, and we had read most of the kid’s books so my mother wrote to the librarian saying I had her permission to read any book in the library and the judgement to evaluate what I read, so I was issued an adult library card. I liked murder mysteries and historical novels. The science section was minimal, but we thought it was great. We liked
“-ologies,” and being kid-like, adopted a new “-ology” every month — ornithology, geology, ichthyology, herpetology, microbiology, etc. Actually, I think we liked the importance of the long scientific words as much as the science itself.
In high school, I did the usual high school activities — dated, danced, hung out with friends, followed the big bands, worried about my appearance and being popular, and tried not to be labelled a “brain.” I always wanted to go to college, but didn’t think it was possible, as no one in the family had that kind of money. However, in my junior year, I decided I was going to go somehow or other and started looking for scholarships.
There were almost no jobs for women in those days and what there were paid very little, so I worked at all the jobs I could get — babysitter, car hop, clerk in the local department store, copywriter for radio ads, china decorator in a local glass factory, seamstress making slipcovers and drapes, and so on.
I was lucky enough to be awarded a full scholarship in advertising design at the University of Cincinnati, a co-op school where one could work a semester and study a semester, and thus work one’s way through. I was so excited and took a train to Cincinnati to accept the scholarship only to find out that you could not co-op your freshman year. I didn’t have the money for room and board, so I had to give up the scholarship. This was a very low point, but I was still determined to be the first person in my family to go to college, so I enrolled in West Liberty State College, which was close enough for me to commute and live at home.
I was working two jobs and trying to go to college, which obviously was not going to work for very long. In my second semester, the college librarian offered me a staff position as assistant librarian, which allowed me to work full time at a reasonable salary. We worked nights in the library, so I had time off during the day to take classes. The art department was minimal, but the chemistry department was reasonably good, so I majored in chemistry.
I did graduate work at Purdue in biochemistry and completed course work for a PhD. However, I was worn down from living on $130 a month and not happy with opportunities for women in biochemistry at the time. So I decided to work a year and make some money to proceed, I was amazed to be wined, dined, and Sweet Adelined at the American Chemical Society job fair in Miami, where I had more than twenty job offers. Sputnik had just gone up and the US was frantic for scientists of any stripe — even women — to play catch-up with the Russians. I accepted a job at Chemical Abstracts, which in those days was on the Ohio State campus.
Newnham: What did you do there and what prompted the job change to SRI? What was SRI like back then?
Feinler: The job at Chemical Abstracts [CA] was an information job. I was assistant editor for microbiology, botany, sugars, gums, and polysaccharides for the Fifth Decennial Index. This was a huge information undertaking and piqued my interest in handling large data projects. We did the indexing by reading the reprint or its abstract and dictating appropriate index entries, which were then transcribed by secretaries onto 3x5 index cards, re-reviewed by us, and stored in order in vaults before being sent to printers for printing. None of the process was computerized. CA was just thinking about automating at the time, but neither the right software nor computers were available yet.
As a single person, Columbus, Ohio, wasn’t somewhere I wanted to spend the rest of my life. We jokingly said Columbus was “centrally located in the middle of nowhere.” I read in Chemical and Engineering News about a group being created at SRI that paired information specialists, in my case chemistry, with research teams. The specialists searched the technical literature, then gathered, and organized the information pertinent to the research team’s needs.
This appealed to me and California seemed like a happening place, so I wrote and asked them to hire me. They wrote back to say they were interested, but did not have a job at the time. It was a long shot, so I forgot about it and decided to go off on a grand tour of Europe with my roommate. Right before we left for Europe though, I received an offer from SRI, which I accepted starting a week after we returned from our trip. When I first got to California, I couldn’t remember where I was, because I had been in so many strange cities in such a short time.
SRI is a non-profit research establishment frequently referred to as a think tank. When I arrived in 1960 it was, and still is, involved in all kinds of research projects and was a very exciting place to work. It did contract research for the US government and private companies, and even for other countries. Projects were typically a couple of years long — some shorter and some much longer — so workers often needed specific information in a hurry. Our group was a service group i.e. we were farmed out to projects because we knew the literature sources and could quickly collect and organize whatever was needed. None of this was done with computers yet, it was all done by brute force by collecting reprints and written material and condensing the information onto myriad numbers of 3x5 index cards, or yellow-lining the reprints, or whatever format suited the researcher’s needs. My first project was a physics project producing a compilation of charge transfer reactions, about which I knew nothing. However, the research group gave me a quick tutorial and we published the compilation — and I got my sea legs at SRI.
Newnham: What led to you working with Doug Engelbart’s team?
Feinler: I moved up to be manager of my service group and as such worked on bigger and bigger information projects, such as the Handbook of Psychopharmacology for the National Institutes of Health, and preservation of biological specimens for NASA Skylab. On one project on fire retardant chemicals, I decided we needed computer help, so worked with two programmers to develop a bibliographic searching system on a General Electric timeshared computer. Doug was anticipating bidding on the Defense Advanced Research Projects Agency [DARPA]
Network Information Center [NIC] project for the ARPANET [forerunner to the Internet], so he dropped by to see what I was doing. He had developed NLS [oN-Line System], which was an amazing system at the time of the late 1960s because it had almost everything we do today on computers — programming, text editing, mixed text and graphics, remote teleconferencing, debugging, hypertext, font selection, document production, the mouse pointer, and on and on. He wanted to use this system as the basis for the ARPANET NIC project.
I saw his system as a great breakthrough for using computers to handle information across networks, so asked him to hire me. He said he didn’t have a job for me at the time, so I dismissed the idea. A few months later he appeared in my office and said he had a job for me now, and asked me if I wanted to transfer to his group. I asked him what the job was, and he said he wanted me to produce the ARPANET Resource Handbook. I asked him what a resource handbook was, and he said he didn’t exactly know, but he needed one in six weeks for DARPA’s demo of the ARPANET in Washington, DC. So that is how I joined Doug’s group.
Newnham: What did you think of NLS the first time you sat in front of it?
Feinler: That I would crash the system. After I learned to use it, I thought it was a very useful and creative tool, with everything right at my fingertips. It had text editing, programming. debugging, a journal system, a database system, email, a consistent user interface, searching, hypertext links, teleconferencing, mixed text and graphics, the mouse pointer, and on and on. Some features were experimental at the time, while others we used on a daily basis. It did most of the things we do today. However, the software was unfortunately way ahead of the existing hardware at the time.
Newnham: Going back to NIC. What was its purpose?
Feinler: The NIC was primarily an information hub for the ARPANET/Internet. At first, in the early 1970s, it dealt mostly with paper distribution, as there were no protocols for file transfer or email. Once email, the first “killer app,” was developed then most interactions were done by email or information servers.
The NIC published the ARPANET/DDN Directory [essentially a phone number and email address book of network users and personnel], the ARPANET Resource Handbook [a directory of computers and resources available for use across the ARPANET], and the DoD Protocol Handbook [a compendium of the protocols or technology upon which the Internet was built].
Over time, we provided a twelve-hour-a-day telephone hotline, maintained the official host table, developed the top-level domain naming system, ran an audit trail and billing system for the DDN, administered the telephone access to the DDN [TACACS], maintained an online and hardcopy library — both at SRI and at the Defense Communications Agency [DCA] in Washington, provided the first netwide WHOIS name server, were the official distributers of the Requests for Comments [RFCs, a network technical note series], ran a large computer facility, and coordinated various official working groups, such as the Network Technical Liaison and Host Administrators. It was through these groups [Liaison and Host Administrators, as well as the Network Technical Working Group] that we all kept in touch. The NIC relayed info and problems out to these groups. And they kept us informed of happenings and problems at their sites or in general. The NIC also made announcements on behalf of our sponsors to the user community.
Newnham: Can you tell me more about your role within NIC and how it evolved over time?
Feinler: I joined Doug’s group in 1972 as an information scientist to produce the ARPANET Resource Handbook, a compilation of computer and program resources available at various government and university centers that were accessible over the ARPANET.
There was a funding hiatus in 1972 and funds were cut drastically. Many of the NIC staff were laid off or left SRI. Shortly after that in 1974, I became the Principal Investigator for the NIC project, and the NIC became a separate project from Doug’s NLS development work. At that time the NIC had a budget of $100,000 and a staff of two or three people including me. Trying to keep the project alive for $100,000 was really hard. Everyone worked many more hours than they were paid, and trying to run the NIC from California on a ten percent pie-slice on a BBN computer in Boston almost caused me to resign. The successful cutover to the TCP/IP protocols was probably the most significant milestone.
Doug’s group was sold off to Tymshare Corp. in 1977, but I remained at SRI to run the NIC and to evolve its networked methods of information handling. DCA took over the day-to-day operation of the ARPANET, and after 1984, the ARPANET became one arm of the much larger Defense Data Network [DDN]. The scope of the NIC project expanded considerably, and I moved up to be director of the Network Information Systems Center at SRI.
When I left SRI in 1989 the NIC had a staff of more than forty, we ran a computer center with a DEC-20, Foonly [DEC-10 lookalike], and Sun servers, and had a budget of about 11 million dollars. This was a very big project at the time, especially for a woman.
Newnham: Speaking of that. What was the ratio of men to women back then?
Feinler: Almost all programmers were male. Although I wasn’t a programmer, so I was usually accepted. Maybe it was my nickname “Jake.” People didn’t always know whether I was male or female.
There were several women working on the NIC project as we grew. Many worked their way up, so it was hard to get them proper recognition. I made them task managers and gave them substantial budgets to manage, as well as responsibility for the task. It was hard to ignore a manager with responsibility for a million dollar task — male or female. Sometimes I had to find out on the sly what men were making for a given job, then insist a woman doing the same job get equal billing and pay.
Newnham: Going back to the NIC. What were some of the major challenges you faced and how did you overcome them?
Feinler: I would say the greatest challenge was too many users and too little computer capacity. Another challenge was that the ARPANET/Internet was still being developed, as it was fast becoming a valued user utility, and the two purposes were not always compatible. At first, computers were huge in physical size, expensive to maintain, but small in actual computer capacity, and in most cases their use was spread very thin in many directions.
Many host computers on the early ARPANET/Internet were being used for other purposes than just network research, so it was difficult for their staff to pay allegiance to all the administrative demands. For example, the NIC maintained the official host name table for the ARPANET/Internet. We could ask a site to implement something important to the ARPANET/Internet, but it might not respond. Some sites did not install a complete host table on their computers, because it took up too much space or was too much work. Consequently, some email could not be delivered, and the NIC got complaints. This is one example of many similar problems.
The NIC resolved these problems by working together with local site staff, and technical gurus, when needed. In this case we developed an automatic server to update site host tables more easily, and pointed out the advantages of having a complete host table installed. Later we worked with network gurus to switch over to the domain naming system in place today.
By the time I left SRI in 1989, computers were small in size and much greater in compute power, but capacity was still a problem because the network had grown exponentially. Today capacity is obviously no longer a problem.
Newnham: How did the inevitable spurts of growth and change affect you and what you were doing?
Feinler: The NIC grew as the network grew. At first, there was myself and one other person, and no programming support. We had to ask Doug’s programmers for help, largely as a favour, to get anything fixed or changed. In 1974, I was able to hire a programmer and the NIC began divorcing itself from using Doug’s NLS system, other than for text editing and some of its database programming.
At that time we began offering many more services via network information servers, as these were more efficient and did not require a login or prior training. We did not have a dedicated computer and had to work in a sometimes minuscule pie-slice on a rented computer.
In 1984, DoD officially adopted the TCP/IP protocol suite and the NIC project expanded greatly. At that time, we obtained the DEC-20 and Sun servers, began transition of our software to the C programming language, began implementing the domain naming system, added many new features to the WHOIS server, and began collecting the audit trail and billing data for the DDN.
Also, by this time almost all network services except the telephone hotline were provided online. In addition we provided written documentation on request or by subscription to users all over the world, who were not on the Internet, but were interested in the packet-switched technology.
An ideal size for an informal research group at SRI is, in my observation, about twenty people. After that more structure is needed, and this was true of the NIC group. Our contract was laid out into specific tasks, so I appointed task leaders for each task, added a contract administrator to help track expenditures and deliverables, and hired a manager for our expanding computer facility. When I left in 1989, we were a group of more than forty people.
Newnham: What was it like working with Doug?
Feinler: Doug was my mentor and a friend for many years, until he passed away. He was a quiet, thoughtful person with a mission and a passion to augment human intellect through a symbiosis with computers. He was all ears if you wanted to talk ideas, but could become a turtle in its shell if you just wanted to complain, and there was lots to complain about early on.
I remember going to his office once with a veritable litany of woes. The system was down, computer was slow, I couldn’t get time on a workstation, no programmer was available to fix things, and on and on. I was on a roll. He listened until I finished my rant, and then quietly said, “What are you going to do about these problems?” I got that he didn’t hire me to bitch. He hired me to think my way to a solution.
Newnham: Can you tell me anymore about the systems that you were using during this time?
Feinler: We started out with Doug’s NLS system, which had many great features, but was way ahead of the computer capacity needed to deliver it to a whole network of users. At first, the NIC project had a pie-slice on a crowded DEC-10 and then a DEC-20 — both at SRI. Then we were switched to Tymshare machines, then to a machine at Bolt Beranek and Newman [BBN] in Boston — which was virtually impossible — then to a purchased in-house Tymshare Foonly, and finally, to a purchased DEC-20 running TOPS 20. When I left, we were converting some activities to SUN servers, but still had the other two machines. We used a variety of programming languages — mostly L-10 and C or C++. We wrote our own DBMS and programs for many of the NIC servers. The domain naming system was adapted from Paul Mockapetris’ Jeeves program.
Newnham: You were working at SRI at the dawn of the Internet. What was it used for in the early days and how did those working to create it think it would evolve?
Feinler: Computers were very expensive when the ARPANET/Internet began. Sputnik had gone up and the Cold War was in full swing. At first it was anticipated the ARPANET/Internet would be used for resource sharing. Scientists would be able to access large computers in real time across the network, and not have to travel to a distant site to use its computer. When email was developed — along with text editing, file transfer, and computer printing — everyone started using it for communication, document production, and text manipulation.
Meanwhile, packet-switching as a networking technology was proving to be superior, so funding moved from resource sharing to advanced protocol development. The ARPANET/Internet was unusual in that it had a rather robust user community at the same time protocol development was going on. The users and the developers often clashed in their priorities.
At first the network was strictly funded by the military. Then other government agencies began to use it. Other networks like BITNET and UUCP [Unix-to-Unix Copy] net found ways to exchange email with the ARPANET/Internet, then finally NSFnet was created and countless students were using the network. Meanwhile, personal computers were flourishing, LAN protocols were developed, the web was developing, the world was wanting and/or getting on, and the government was still footing the bill. Thus, the decision to allow commercial traffic and let the Internet pay its own way.
Newnham: I understand security/privacy wasn’t such an issue back then, but what were issues? What were the things people worried about early on when it came to the Internet?
Feinler: Speed and congestion were major concerns. Protocol changes, especially the switch to TCP/IP — a major technological event — meant sites had to change a great deal of software. Some didn’t make it and left the net. Users were often caught in the middle and couldn’t get their day-to-day work done. At the beginning, developers were the main users, and they wanted transparency so they could build and test things out across the whole net and work freely back and forth. Then, many students crept onto the net one way or another, some legitimate, others because it was the best game in town, so security was kind of added on rather than built in. “Hackers” in those days were apt to hack into a system, find a flaw, and tell the developer as a favour so they developer could fix the problem. Few were trouble makers, they just wanted on.
When personal computers began being used for office work and before LAN protocols, users still had to send and receive email on host computers, so these were hopelessly bogged down. Flaming was a problem, but flamers were often jumped on by others and told to “knock it off.” If the flamer was part of an online working group and didn’t stop flaming after a warning, they could be removed from the group. Many email systems used different conventions and user terminology, so getting email from one network to another was a challenge.
Newnham: How did your work evolve as the tech evolved?
Feinler: Mostly we went from paper and hardcopy documentation to doing most everything online. Also, the transition away from NLS and the transition to TCP/IP were a challenge. We built our own software based on the premise that NIC users were episodic users — i.e., they did not have the time or inclination to spend learning how to use what we offered, so it had better be obvious to start with. Needless to say keeping up with the email and ever expanding user community was never ending and exhausting.
Newnham: I have heard from others that the real activity online happened at night. What was it like in those days?
Feinler: It was great fun. The computers were so wedged during the day that night was the only time any real work could get done. Night users were usually students working on their theses or programmers trying to run programs along with whatever kid hackers that could worm their way on. We “chatted” back and forth by typing, and took breaks to talk and exchange network gossip and ideas.
I once travelled to MIT where I had a Chinese banquet with about fifteen people who were network friends, but whom I had never met in person. It was mind boggling! I sometimes did a rollover where I worked all night then stayed to work the next day to cover phones and daytime happenings. The hours of 4 to 5am were the worst. I drank a lot of Cokes and Mountain Dew.
So it was exhilarating and frustrating at the same time. Everyone wanted the network to succeed, so there was great esprit de corps, but things often broke, or were painfully slow, or mismatched so that meeting a deadline was a gamble. We all worked very hard though and took pride in what was evolving.
Newnham: What turned out to be the most surprising discovery during this time?
Feinler: I thought the sociology of the network was even more interesting than the technology. The Internet completely changed the way people communicate and perform knowledge work, and the web has put the world’s knowledge at everyone’s fingertips. Who would have imagined terabytes of capacity on an iPhone?
Newnham: You left SRI in 1989 and contracted for NASA. Can you tell me about why you left and the projects that you worked on for NASA?
Feinler: I was burned out. I had just renegotiated the contract, so the group was funded. I wanted to do something else, or actually do nothing at all for a while. I had been invited to be a visiting scholar in Brazil and I was going to do that and travel around South America for a while. Unfortunately, the day I was in Washington introducing our clients to my successor, Dr. Frank Kuo, my stepfather had a heart attack and died. My mother was partially blind and could not live alone in Florida, so instead I spent the next year or so, primarily focused on family and personal matters.
I attended an IETF meeting at NASA Ames and ran into Bill Jones, a former technical liaison. He asked me what I was doing, and I said nothing, so he invited me to come to work for the NASA Science Internet [NSI], which I did. I did a variety of things there. I served as a requirements manager for NASA scientists based at the large ground-based telescopes, such as Aricebo and Ceretola. I helped transfer the NASA NIC from DECnet at Goddard to the Internet at NASA Ames. I helped establish ground rules for managing a website at NASA. I wrote proposals and so on.
Newnham: Looking back, what part of your career are you most proud of and why?
Feinler: I am proud to have had a small role in the development of the Internet, a technical phenomenon that has changed the way the world learns and communicates. I also think the conversion of the whole network to the TCP/IP protocol suite remains impressive. This was a monumental technical undertaking in which we participated.
I am glad to have had the opportunity to foster careers for several people in my group, who started from the ground up. I always expected people to do their best, and they almost always did. Several went on to do very well. I was proud that I gave them their start on the NIC project.
Newnham: What do you think about the way the Internet went? Did you see the tipping point coming?
Feinler: Well I knew that the government was not going to continue paying for all of it, and that the web was another “killer app” that changed everything. It was too useful to do away with, so commercialization was an obvious solution. I did not anticipate that naming and addressing would become a multimillion dollar business though.
I hope the Internet will maintain net neutrality. My fears are that porn, cyber espionage and crime will bring it down. We need a body of international law worthy of the task of controlling these. I worry about the gap between those who have access versus those who do not. And I also worry about people’s ability to separate facts from opinions, and recognize that the two are not the same thing.
Newnham: Finally, what do you wish your legacy to be?
Feinler: I never thought of myself as important enough to leave a legacy. I hope people consider me a person of integrity. I am proud of the work we did and the career paths that I was able to create for young people, particularly women.