The rise and fall of the Gopher protocol
It was mid-March 1992, and Mark McCahill had never been to San
Diego before. Back home in Minneapolis, the skies had been
dumping snow for six months, and would keep at it for several
more weeks. McCahill checked into the Hyatt Islandia, an 18-story
high-rise hotel overlooking Mission Bay. There were palm trees,
he recalls. Boy, was it nice.
McCahill was then in his mid-30s and managing the Microcomputer
centre at the University of Minnesota Twin Cities, which
facilitated the emerging use of personal computers on campus. He
and Farhad Anklesaria, a programmer in the centre, had been
invited to address the 23rd Internet Engineering Task Force
(IETF), an elite convocation of academics and government
officials from around the world who were literally deciding how
the internet should work.
The gods of the internet, McCahill says, though in other circles
they would have gone unnoticed. In his memoirs, another internet
pioneer, Tim Berners-Lee, describes these gatherings as people
in T-shirts and jeans, and at times no footwear. They would meet
in different small rooms and talk excitedly.
The internet then, as now, was a vast array of information stored
in random computers around the world, only there was no easy or
consistent access. It was difficult even to discover what was out
there - there were no good search engines. The most popular
protocol, or method of retrieving information from another
computer, was FTP (file transfer protocol), the primitive,
labour-intensive equivalent of knocking on someone's door and
asking if you could carry away his piano.
The IETF had been convening since 1986 to iron out these issues,
which had prevented the internet from becoming the Intergalactic
Network its originators had foreseen, instead remaining the
limited domain of physicists and the military. But this meeting
felt different. For the first time, the internet seemed on the
verge of going public.
On March 18, in a conference room of the hotel, Berners-Lee
presented one possible breakthrough: the World Wide Web. It was
evening. Many of the 530 conference attendees had already gone to
the bar or to dinner. To the curious who stayed behind,
Berners-Lee explained that the Web could be used to connect all
the information on the internet through hyperlinks. You could
click on a word or a phrase in a document and immediately
retrieve a related document, click again on a phrase
in that document, and so on. It acted like a web laid over the
internet, so you could spider from one source of information to
another on nearly invisible threads.
Two other programs with the potential to expand access to the
internet WAIS and Prospero were discussed in the same
session. In the reports of people who saw the presentation, the
Web did not come across as the best of them, or even as
particularly promising.
The next day, in the light of the afternoon, McCahill and
Anklesaria presented the Internet Gopher. It was simple enough to
explain: With minimal computer knowledge, you could download an
interface the Gopher and begin searching the internet,
retrieving information linked to it from anywhere in the world.
It was like the Web but more straightforward, and it was already
working.
In fact, most attendees needed little introduction to Gopher the
software had been out for months. It was the developers they were
curious about, the Minnesotans who had created the first popular
means of accessing the internet. People wed never met were
telling us how they were using our stuff and adding things to
it, McCahill says. We had no idea how big Gopher was going to
be until we experienced this firsthand and realised that growth
could be exponential for a while.
In the years that followed, the future seemed obvious. The number
of Gopher users expanded at orders of magnitude more than the
World Wide Web. Gopher developers held gatherings around the
country, called GopherCons, and issued a Gopher T-shirt worn by
MTV veejay Adam Curry when he announced the networks Gopher
site. The White House revealed its Gopher site on Good Morning
America. In the race to rule the internet, one observer noted,
Gopher seems to have won out.
McCahill's father was an executive for Conoco, the oil company,
which moved him around the country about every two years.
McCahill was in junior high when the family finally settled in
the Twin Cities. He graduated from the U in 1979 with a BA in
chemistry, spent a year studying effluent in rivers, and realised
he liked the computer analysis (heavy number smashing) more
than the chemistry itself (kind of dirty). So he took a job in
the Us Microcomputer centre, programming some of Apples first
personal computers.
The Twin Cities were a proto-Silicon Valley then, with a long
history of producing some of the worlds most powerful computers
at UNIVAC, Control Data Corporation, and Engineering Research
Associates, which supported the work of Honeywell, IBM, and other
local tech firms. The Minnesota Educational Computing Consortium,
or MECC, formed in 1973 to get computers in schools and created
software for them most famously The Oregon Trail. By the early
1980s, when only a fraction of schools in neighbouring states had
computers, there were about three or four in every public school
in Minnesota.
When McCahill began working in the Microcomputer centre, a turf
battle was heating up a religious war, McCahill calls it
between the high priests of computing who oversaw the Us
venerable mainframes, the enormous machines that once occupied
entire rooms, and the growing cadre of personal-computer
converts. Microcomputer guys were as far out of the mainstream
as you could get and still be a part of the Us computer centre,
McCahill says. Anklesaria, who had earned a doctorate in genetics
before gravitating to computer science, says the Microcomputer
centre was a splinter group when he joined in late 1987. The
mainframe was still the only thing the Mac was considered a
toy.
McCahill sided with PCs. The idea of democratizing access to
computing, putting computers in the hands of everyday people
that resonated with me, and that was part of all the early PC
stuff, he says. If you were interested in PCs, you just
absorbed that attitude by osmosis. It was in the air and the
water.
McCahill, who had long hair then that he now pulls into a
ponytail, spent his free time wind-surfing on Lake Calhoun, and
says the PC revolution looked like a good wave to ride it
would enable me to do what Ive tried to do ever since: take
technology that is cutting edge and get it to the point that its
palatable to mom and dad and English majors.
In the late 1980s, McCahill and Anklesaria developed the first
popular internet email system, called POPMail. It was partly for
selfish reasons, McCahill says. I wanted more people using
email so I didn't have to walk down the hall to my mailbox to
collect my phone messages on little slips of paper. Instead,
secretaries could send an email. For that matter, so could
callers.
At the same time, the U was determined to network its computers on
the internet in a so-called campus-wide information system, or
CWIS, and the schism was delaying development. By early 1991, a
committee of more than 20 department heads and computing
specialists had been meeting for months, producing a long list of
demands including the use of mainframe computers but zero
code. They had some complicated shit for doing searches that I
didn't want to do, McCahill says. Farhad didn't want to write
it at all.
But I had to show something, Anklesaria says. So he stripped the
program down to its simplest parts a basic protocol for making
information in one place available somewhere else. He cobbled it
together on a Mac, writing a server (a program enabling a
computer to serve up requested files) and a client (how most of
our computers are programmed, enabled to search for and request
those files). And I said, This stuff kind of works. Since we
had nothing else, we went for it.
McCahill pushed for a full-text search engine something we now
take for granted and borrowed the gist of one from a computer
system called NeXT, which had recently been invented by Steve
Jobs. We had this marriage of Farhads super-simple protocol for
saying give me a list of items, a menu, says McCahill, and my
thing of having a way of searching, and we glued those two
together.
It was plain text no pictures, given that modem speeds were so
slow. And it was organized like the one information source most
people alive in the early 1990s were familiar with: a library,
with similar subjects grouped together. You just pointed your
gopher, as the lingo went, to any Gopher site you wanted to
explore, and there you were, burrowing through the internet. It
was so simple that just about anyone could make it work, even an
English major.
It was one of the rare times when we both looked at each other
and said, Holy shit, weve got a really good idea here,
McCahill recalls. Anklesaria called it the Internet Gopher, a
triple play on words: the Us mascot, a critter that digs, and a
go-fer one who fetches. We figured that if we called it
Gopher, the committee couldnt complain, McCahill says. Its
the school mascot!
Sir Tim Berners-Lee he was knighted in 2004 was born in London
in 1955, less than a year before McCahill. His parents were both
mathematicians. While McCahill's father was being shipped from
Colorado to Oklahoma to Minnesota, Sir Tims parents were
developing the worlds first commercially available
computer.
Berners-Lee grew up with the internet, or at least the concept of
it. As a child, he built mock computers out of cardboard boxes.
He came home from high school one day to find his father writing
a speech on how computers might someday make intuitive
connections, linking information the way the brain uses random
associations to link thoughts.
He was a teenager when the internet began as the ARPAnet,
connecting a handful of computers at Stanford and other
universities. It was designed as a defense, a secure means of
communication should the Soviets destroy the American telephone
system, though its practical purpose was to allow scientists to
use the computing power at another facility for massive
calculations.
Berners-Lee earned a degree in physics from Oxford where he
built his first computer with parts from an old television, a
calculator, an electronics kit, and a car battery then worked
as a software engineer for a few years before taking a job with
CERN, the famous particle-physics lab in Geneva, Switzerland. By
then, in 1984, the internet had gone global; CERN was the largest
internet node in Europe.
Yet using the internet was problematic even within the CERN lab.
When you wanted information, Berners-Lee later recalled, often
it was just easier to go ask people while they were having
coffee. He decided he could do better. He had never forgotten
his fathers research on the brain, and when he wove the Web, it
was based on this holistic, serendipitous, strangely rewarding
experience of surfing from one vaguely related idea to the
next.
He finished a model of the Web in 1989, but for years it went
nowhere. The concept was too abstract and it only worked on NeXT
computers. But in early 1991, just as McCahill and Anklesaria
were conceiving the Internet Gopher in Minnesota, the first Web
servers outside of CERN were switched on.
After outlining the Gopher protocol, McCahill and Anklesaria
pulled together four programmers who worked in the Microcomputer
centre to write the software. McCahill wanted it in a hurry. Not
because he was racing Berners-Lee, but because he wanted it over
with. He wanted it before the next CWIS committee meeting, in a
month.
The programmers were young guys, mostly in their 20s and, like
McCahill, mostly huge Nirvana fans. Paul Lindner, a coding
wunderkind from northern Minnesota who was dubbed the Gopher Dude
for his evangelism, had long metal-head hair and signed Gopher
emails with lyrics like You have to spit to see the shine from
Babes in Toyland. Early Gopher servers were named Mudhoney,
Danzig, and Anthrax. The sole outlier in the microcomputer mosh
pit was Bob Alberti, a programmer who named a server Indigo, as
in the Indigo Girls.
The centre was in Shepherd Labs, a hulking cement building built
like a tank in 1968 on the Us Minneapolis campus, with concrete
floors and no windows. Early on, it was used for NASA materials
research. There were pipes with strange fluids running through
them, Lindner recalls.
The microcomputer team, in addition to developing software for the
U, ran a showroom for students and faculty interested in buying a
Mac, taught computer training classes, tested software, and
served as a help centre for people with PC problems walk-in and
call-in. Everyone would answer phones at least one day a week,
Lindner says, even if you were programming. That way you were
close to the pain you were inflicting on people if programmers
today still took calls, wed have more user-friendly
software.
Gopher, however, was claiming more and more of their time. It
became infectious, Lindner says. How can we build this into the
science fiction of our dreams, access to all the information in
the world, the library of everything
The team, in 36-hour sessions fueled by beer, pizza, and speed
metal, finished writing Gopher in about three weeks. They
installed the first computer running a Gopher server a Mac
SE/30, a little droid of a computer with an iPad-size monitor
built in in a narrow hallway between their offices and the
showroom, in a closet with metal shelves. It became known as the
Mother Gopher.
The committee meeting where the team first presented the Gopher
protocol was a disaster, literally the worst meeting Ive ever
seen, says Alberti. I still remember a woman in pumps jumping
up and down and shouting, You cant do that!
Among the teams offenses: Gopher didn't use a mainframe computer
and its server-client setup empowered anyone with a PC, not a
central authority. While it did everything the U required and
then some, to the committee it felt like a middle finger. Youre
not supposed to have written this! Alberti says of the groups
reaction. This is some lark, never do this again! The Gopher
team was forbidden from further work on the protocol.
After the meeting, McCahill leaned on the director of the computer
centre, a Chinese-American man named Shih Pau Yen, who had
supported Gopher all along. I said I would quit before I stopped
working on the coolest thing wed ever created, McCahill says.
Yen ran interference, and the Gopher team kept working on it in
their own time.
In this bureaucracy of little fiefdoms, where everyone had a hard
time working together, the one thing that could unify us was
telling everyone else off, Lindner says. That was our rallying
cry.
Finally, in April 1991, still unable to persuade the U to take on
Gopher, Lindner released it into the wild. He made the Gopher
software available via FTP the most popular way to share
information on the internet at the time and wrote a brief,
quiet announcement on an internet mailing list: Hey, weve got
this thing, come and get it.
Within months, the team was hearing from Gopher users around the
country. It was the first viral software, Alberti says. All
these people started calling the U and pestering the president
and other administrators, saying, This Gopher thing is great,
when are you going to release a new version And the
administrators said, What are you talking about
The ban was lifted. Never numbering more than six core members,
the team fanned out to conferences to spread the Gopher gospel,
while continuing to improve and diversify the protocol. In the
process they ended up laying the foundation for much of how we
navigate the internet. The first hyperlinks. The first bookmarks.
McCahill, thinking of windsurfing, even coined the term surf the
internet.
A video of veejay Adam Curry wearing an Internet Gopher World
Tour T-shirt on MTV.
Within a year, there were hundreds of Gopher servers. Berners-Lee,
who had publicly introduced the World Wide Web a few months after
Gophers debut, used Gopher to do it. People look at the World
Wide Web today and think it sprang out of Tim Berners-Lees
forehead, Alberti says. But the fact is, the only way he was
able to spread the word about the Web is because the Internet
Gopher was there to allow people to download his files, find a
discussion group, and talk about it.
We had the right product at the right time, McCahill says.
People were looking to expand the internet beyond physicists
stuff. Gopher could do that. It was simple to use, it could
network lots and lots of computers. It gave people a reason to
say, hey, this internet is good.
Lindner was solicited for side gigs twice he went to Ecuador to
set up Gopher for the countrys fledgling internet. Thats when
I knew we were really onto something, he says, when I was
helping wire a whole country.
Al Gore, then a U.S. senator, came to visit. Four GopherCons, held
between 1992 and 1995, drew reps from the New York Times, the
World Bank, Microsoft, and other global heavyweights. The Gopher
T-shirt, black and scribbly, listed the names of places with
Gopher servers on the back, in the style of rock tour shirts. It
was an apt metaphor, as Gopher team member Daniel Torrey told the
Pioneer Press in 1996: We thought we were rock stars.
Some team members dreamed of fortune to go with their fame. But
the internet was not yet open for business. It had been built on
dot-mil and dot-edu, on public funds. Programmers shared source
code; if you needed something, someone gave it to you. A dot-com
address was considered crass. It was as though all of TV was
PBS, Lindner says. No commercials.
Still, Alberti raised the profit potential of Gopher with Shih Pau
Yen. Before coming to work in the Microcomputer centre, Alberti
had helped create the first online multi-player role-playing game
called Scepter of Goth, an ancestor of World of Warcraft and
the like. I said we should take this private, we should make a
business of this and make some money off it, Alberti recalls.
He looked at me like Id just grown another head.
Eventually, though, the U did want some money for itself. At
GopherCon 93, Yen announced that for-profit Gopher users would
need to pay the U a licensing fee: hundreds or thousands of
dollars, depending on the size and nature of their business. Many
users felt betrayed. In the open-source computing spirit of the
day, they had contributed code to Gopher, helping the team keep
up with the times. Now they were being asked to pony
up.
The reaction deflated the team. As hard as they were expected to
work on Gopher, they were never relieved of other duties,
including answering the Us help line. We never got additional
funding, we were going broke, Alberti says. We had the whole
internet yelling at us, when are you going to update your
software, when are you going to put images on Gopher pages, and
make this smoother and better And were like: Were six
guys!
For a while, the U threatened to get rid of them altogether, a bid
to outsource the universitys computer work, provoking one
programmer to bug a computer in Morrill Hall, the Us
administrative centre, so the team could listen in on
discussions.
Asking for a contribution seemed reasonable. When it backfired,
the team posted a defensive letter to Gopher users in March 1993:
There has been a lot of hysteria, misinformation, and rumor
floating around. In a time where we are having budgets slashed,
it is impossible to justify continued (increasing) resources
being allocated to Gopher development unless some good things
result for the University of Minnesota. This is a fact of life.
Before you go off and flame once more, ask yourself if you want
to get YOUR particular server going with as little fuss and
expense as possible or if you just want to stir up the
soup.
That socially killed Gopher, Alberti says of the licensing fiasco.
Yet it wasnt the end. In 1993, Gopher was still far more popular
than the World Wide Web, and Gopher traffic grew by 997 percent.
But the Web was starting to catch up that year, it grew by
341,634 percent.
At the San Diego internet conference, in 1992, Berners-Lee had
pulled McCahill and Anklesaria aside on the last day of meetings
and asked if they wanted to collabourate on a Web/Gopher hybrid.
Some melding of their different designs an internet
super-system though it wasnt clear how that might
work.
Tim is a great guy, but hes a little odd, a little scattered,
McCahill says. Talking to him is like a ball of twine
experience. The Web you see now, thats how he thinks.
McCahill told Berners-Lee that he would need to look at the Web
more closely. But it wasnt much to look at when McCahill went
back to Minnesota and examined it. There were no graphics yet. It
was still only running on NeXT computers. I wasnt feeling it,
McCahill says. I told him, Tim, I dont think so. Of course, I
look back and say, I might have been wrong.
Soon enough, the Web did have pictures and was available on more
platforms. In 1993, the first popular Web browser, Mosaic, was
introduced for sale, breaking the commercial taboo of the
internet and suggesting to McCahill at least that tech
investors had taken sides. The fix was in, he says.
In 1994, modem speeds doubled, and the interminable rendering of
images on the Web once dubbed the World Wide Wait greatly
accelerated. PCs began to be sold with these faster modems built
in. To anyone looking for a simple, even crude explanation for
the Webs rise, this is it: the ability to view a reasonable
facsimile of a naked woman in the privacy of your own home.
Thats what came to drive a lot of the internet, Alberti says.
Porn.
The Internet Gopher, with his text-only menu and gloss-less,
institutional mien, couldnt keep up. He had fallen off the wave,
and almost overnight was revealed as a buck-toothed square,
ignored by the girls on the beach, his surfboard held together by
duct tape. Obsolesced, as one observer put it in 1994. A
has-been.
I remember the exact moment I knew I was no longer on the right
track, says Lindner. It was September 9, 1993. I was invited to
give a talk about Gopher at Princeton, and I had my slides all
printed up on my little university-budget black-and-white foils.
The person presenting before me was talking about the future of
the Web, with full-color LCD projection. I said, I think I see
where things are going.
Mark McCahill's entry on the University of Minnesotas Wall of
Discovery, which celebrates the accomplishments of students and
faculty.
For McCahill, the realization happened on the street. I saw a URL
on the side of a bus, he says. Thats when I knew the Web was
all about advertising. Gopher was not good for advertising. I
knew it would start winding down.
In the spring of 1994, Web traffic overtook Gopher traffic for the
first time. Gopher, within a few months, began to decline. Its
reign had lasted three years.
The coup de grâce came from the U itself. In 1995, Dr. John
Najarian, a renowned transplant surgeon at the university, was
indicted for tax evasion, embezzlement, and fraud. Although
Najarian was later cleared of all charges, the National
Institutes of Health (NIH) put the U on probational status,
threatening tens of millions of dollars in grants and the
schools viability as a top research institution. It was an all
hands on deck situation, McCahill says.
The issue was essentially simple: bad accounting. Paperwork was
literally on paper, and transaction records sat around for weeks
before making it into a ledger, if they made it in at all. The
Gopher programming team was diverted to creating a more
accountable accounting system for the U, what turned out to be
the first Web-based transaction program. Thats what my kick-ass
development shop did for a year and a half, McCahill says, to
show the NIH that we were cleaning things up. By the time they
finished, the Internet Gopher was dead.
In the beginning, when the Mother Gopher was new and there were no
other Gopher servers to link to, Gopherspace was empty. The
Gopher team, to demonstrate the usefulness of their invention,
dumped in a cookbook and searched for eggplant. As late as 1993,
the most popular information in Gopherspace included recipes,
weather, phone books, and movies of chemical reactions.
Gopher represented a simpler, more naïve time, says one of its
modern-day fans. Others call it a purer way of navigating the
internet, of making structure out of chaos.
Today, there are about 140 Gopher servers still out there, many of
them relatively new. The tech world is not a sentimental place,
but it does appreciate simplicity and irony in equal measure
many of these servers were set up on April Fools
Day.
Farhad Anklesaria: You have to be at the right time and place to
have your technology take off and become popular.
Cameron Kaiser, perhaps the Internet Gophers greatest advocate
these days (he runs a Gopher support site called the Overbite
Project), figures he spends about 25 percent of his internet time
in Gopherspace. Its actually rather nice to have a small
ecosystem because no ones running annoying ads in Gopherspace or
trying to track your browsing habits, he says. The protocol
makes the former hard and the latter almost impossible.
As the Web has become synonymous with the internet, and we conduct
more of our lives online, weve learned to abide the
misinformation, solicitations, and scams that thrive in its chaos
though it has changed us and society in ways we are only
beginning to understand. If Gopher had won out, who knows how
things would be different.
The Gopher team, who might understandably be disappointed, seem
genuinely OK with how things turned out. More than OK, actually,
as though they expected it. They were first, after all. And in
the march of technology, the first are eventually last. This is
the natural order of things, Anklesaria says. You have some
building blocks the natural thing to do is build on them.
Thats how civilization bootstraps and how we progress.
If they are surprised by anything, its that the Web was the
system that surpassed them and has outlasted everything else.
There was a protocol that was better than either the Web or
Gopher, they say, called Hyper-G, developed in Austria, which
never got off the ground. You have to be at the right time and
place to have your technology take off and become popular, says
Anklesaria. If youre at the base of a wave, you have a chance
to rise with it. But if youre already on top of the wave and the
wave has broken, thats not good if youre too early, you wont
last.
Bob Alberti: The only way [Tim Berners-Lee] was able to spread
the word about the Web is because the Internet Gopher was there
to allow people to download his files, find a discussion group,
and talk about it.
On July 1, 1997, or maybe in 2009, most likely in the 2000s,
someone walked into the closet in Shepherd Labs and unplugged the
Mother Gopher. No one can agree on the date. No one from the
Gopher team was around.
While I was at the U, McCahill says, it was a point of pride
that wed keep running the server. But McCahill left in 2007 to
become a systems architect at Duke University, developing
instructional and research computing technology. Lindner left in
1996, having followed a Gopher side gig to Geneva, Switzerland,
where he worked for the United Nations, and is now a software
engineer for Google. Anklesaria retired last spring. Of the core
Gopher team members, Alberti is the only one still at the U, as
an information security architect.
At its peak, the Mother Gopher consisted of 10 Apple IIci
computers. But when it was finally euthanized, who knows what
shape it was in. There was no ceremony. Nothing was carted off to
a museum. Gopherspace simply became emptier, and the world
without the Web became harder to imagine.