Cookies, Chaos and the Browser: Meet Lou Montulli

The #2 aquatic-themed celebrity site on the Web in 1997 belonged to red-suited Baywatch babe Pamela Sue Anderson.

The #1 site featured a live cam mounted on an SGI workstation in the office of a 23 year-old University of Kansas “Jayhawk” named Lou Montulli.

It was pointed at a pod of exotic fish in a 90-gallon, fluorescent-lit acrylic tank. The fishes’ (rather repetitive) antics played out on the “Download” page of Netscape.com, which offered the first widely-adopted portal into this thing called the World Wide Web.

And it says something about the pace of change in our time that Lou Montulli – one of the pioneers of so many foundational elements of the internet – is still only in his 40’s.

Peter Adams / Faces of Open Source

Here are some of the things Montulli engineered or helped prod into life:

• HTTPS
• server push and client pull
• HTTP proxying and proxy authentication
• the cookie
• web forms
• “about:” URL
• <br>line break</br> tag
• <blink>blink</blink> tag
• animated GIFs

Yet Montulli laughs when I ask the obvious question. “No,” he says, “I don’t feel I’m a legend. And there are a lot of people around me who’d agree.”

The most notorious of his inventions is of course the browser cookie. Montulli has mixed emotions about his spawn, although it’s a durable solution to a serious problem with the web.

WHY DID YOU DECIDE TO INVENT THE COOKIE AT NETSCAPE?

It came out of the problem that HTTP is anonymous. People were anonymous all the time. But there was an obvious need sometimes to recognize them when they came back to a site.

The web came out of universities and academics, and we were all pretty distrustful of government. You could call us Libertarians. We didn’t want anyone to be able to track people. So we needed a mechanism that allowed you to be remembered without being tracked.

There wasn’t a lot of impetus at first for anyone to solve this. I finally got around to it because [Netscape’s product team] wanted a shopping cart. If you have a shopping cart and put something into it, obviously you have to recognize the person when they return.

The trouble with HTTP Basic Auth[entication] is it’s ugly. It’s not friendly. You have to create an account up front. We needed something besides that. The cookie was a general mechanism that allowed people to create interesting apps on the web.

SO IT WASN’T FOR ADVERTISING?

No.

HOW DID IT GET TO BE USED FOR ADVERTISING?

Cookies can be used for tracking only if you have the combined efforts of a lot of websites that participate in one giant … [wink] … conspiracy. In order for an advertiser to track someone across sites they need a relationship with all the sites. They use cookies and referrer fields and JavaScript.

Which is of course our beloved ad tech: the labyrinthine mosh pit of ad servers and platforms and data brokers and analytics that arose to do exactly what Montulli describes: track people across multiple sites. The DoubleClick ad server was born in Kevin O’Connor’s basement in Alpharetta, Georgia in 1995.

But the humble remit of Montulli’s original cookie was to allow a single domain to create “sessions” and also recognize a person if they returned to that domain. One popular early proposal in WWW design groups was for a persistent browser ID; Montulli opposed it.

He admits that third-party cookies, which piggyback on embedded page content like image tags, were “a problem that I missed.” And he has written about a fascinating moment in 1996 when he – all by himself – decided the browser would not block third-party cookies by default. In a way, Lou Montulli created ad tech 1.0.

DO YOU THINK THE COOKIE IS DOOMED?

No. It’s fundamental to the web and most apps won’t work without it. There’s been a lot of time and effort spent trying to improve it. But nobody’s come up with anything better. It likely won’t be replaced with something else.

It could be removed from ad tracking. I’d like that so people would stop blaming me for it. [laughs] Advertisers won’t let it die. If it was [outlawed], they would just move to another technology. There are other technologies that would suffice, but they would remove transparency. That would remove the ability for the end user to control it.

Montulli is referring to techniques such as ‘fingerprinting,’ using data such as browser settings and response times to identify a device or person; and methods such as “JavaScript sending secret codes to other browsers” and so on.

WHAT DO YOU PREDICT WILL HAPPEN NEXT?

We’re in an interesting stage right now with advertising and the ability to disclose [tracking], and your ability as a consumer to control it. It’s built up over years of trial and error. Like it or not, if you want to see content for free you are going to have to put up with ads.

You can give up free sites, give up the cookie – and go into the black hole of [secret tracking]. That’s the worst case, I think. We could also legislate so there’s no more personalized advertising. But that would kill advertising on the web. We’d be forced into paid content. These are not easy solutions.

Apple started this war with the Safari browser. A bunch of companies sprung up with alternative [tracking] models. There are a lot of methods that would work. But they’re all worse for the consumer.

An affable, hard-working music- and MAME-enthusiast, Montulli’s lived in Northern California since moving out for Netscape. He spent his childhood following his father around military bases and stayed in Lawrence, Kansas after graduation to work in the University of Kansas computer lab.

He monitored tapes and dot-matrix printers and was promoted “out of the basement” onto the school’s I.T. help desk. It was there he got involved in the proto-web.

HOW DID YOU START GETTING INTERESTED IN WEB BROWSERS?

I was using software called ELM for email on UNIX. Every time you logged out of ELM it printed out the author’s name. I was not a particularly good student. So I thought, if I had a program with my name on it, I’d probably get a better job.

There was this project kicking around to build a campus-wide information system. In those days, we used BBS’s and networking was primitive. I stayed up all night and made this prototype program using Gopher and HyperRez – an open source hypertext reader. Gopher had terrible UI but did networking. So putting the two together was a good solution and became the first version of Lynx.

Lynx was a text-based browser that was widely adopted by universities, in particular. Montulli was working on a graphical version when he saw an announcement that Marc Andreessen and a team at the University of Illinois’ National Center for Supercomputing Applications (NCSA) had released a browser called Mosaic “that did exactly the same thing.”

WHAT DID YOU DO WHEN YOU HEARD ABOUT MOSAIC?

I converted Lynx to HTTP and HTML [protocols]. It became one of the three parts of the web ecosystem. The other parts were Mosaic and [Thomas Bruce of Cornell’s] Cello browser for Windows. By the time I left Kansas we had over a million people using Lynx. It was the most popular browser by number of users.

DID IT MAKE YOU A RICH YOUNG MAN?

[Laughs] No. It was free open source software. The university was kind enough to pay me $8 an hour for 20 hours a week to work on it.

These were the days when HTML and HTTP were in active development. The internet was still an academic exercise. Off-campus connections were made over phone lines and modems. Long distance charges applied. Graphics were a rare luxury. And the foundations for the Web were being laid by a small cadre of workaholic dreamers perched around the globe.

WHAT WAS THE WEB DEVELOPMENT COMMUNITY LIKE IN THOSE DAYS?

There were basically three camps that could move things forward. There was Tim Berners-Lee and the people at CERN — they gave feedback on specs and worked on server software. Then NCSA was developing clients [like Mosaic]. And there was me working on Lynx … There weren’t any standards bodies. There were maybe twenty people involved and only three or four were developing code.

SO HOW DID YOU FINALLY DECIDE TO JOIN NETSCAPE?

Well, Marc [Andreessen] left NCSA and met Jim [Clark, SGI founder and serial entrepreneur] through friends. Marc told Jim about the Web, and he got excited about it. We all met in Illinois – all of the developers – in March of 1994. We started the company with guys from Mosaic, from CERN, and some engineers Jim knew from SGI.

Inspired by Mosaic, Netscape Communications became the fastest-growing company in the history of the world. It soared from a dozen people in 1994 to about 4,000 by 1996. It broke $100 million in revenue within one year. (This story has been told with his usual flair by Michael Lewis.)

Any of us who were sentient in, say, 1995 remember Netscape. It inspired us to buy modems and “log on” to the web for the first time. By 1997, more than one in three U.S. households had a home computer. So it’s surprising to learn that its financial success was a kind of accident.

WHAT WAS NETSCAPE’S BUSINESS MODEL SUPPOSED TO BE?

Basically, we were going to give it away to clients and sell the servers. It was razors and razor blades. We were going to sell enterprise software, because those were the people who can afford it. But as they say, no plan survives contact with the enemy. You could download it for free, but businesses wanted to pay for the client [browser]. You could also buy it shrink-wrapped in a store.

Businesses were unclear about whether they had to pay for it. We were willing to let them think they had to pay. And we got addicted to the “crack” of client revenue. It actually killed the company when we were forced by Microsoft’s aggression to give it all away for free. We got addicted to revenue we weren’t supposed to have. It sucks we didn’t succeed as much as we should have – but we made a big impact.

We basically implemented the foundations for the [Web] app development environment that’s taken over programming today.

WHAT WAS IT LIKE TO BE IN AN ENVIRONMENT LIKE THAT?

It was a very exciting time. We hired almost all the people who were building the Web, and we worked non-stop. Some survey agency called me up at the time and asked me how many hours a week I worked. I said 120. The fields [in the survey] only went up to 99. I only slept every other day for ten hours.

And it was good software. It was stable, fast and looked good for the time. Functional software … needs to be seamless, beautiful, something the user wants to use. We took over the market within months because we created the most stable, elegant application. We were way bigger than Word and everything else at the time.

BUT WHY THE ‘FISH CAM’?

I had been an a aquarist at college. I basically spent all my waking hours in the office. So I brought my fish tank in. My SGI Unix workstation had a very rare camera on it — the first of its kind. SGI wanted to do videoconferencing. And I saw the Coffee Cam [at Cambridge] and decided to do a version. We used it as a test bed for a lot of advanced features.

WHICH FEATURES?

The idea of images refreshing in the background. That led to animated GIFs. The first animated GIFs ever were of the fish. Dynamic HTML was first tested there. The ability to move HTML elements around on the screen. The early web was very static.

Netscape became the cynosure of the Dot Com Bubble, going public in 1995 – it’s said – so Jim Clark could build a yacht. The stock quadrupled on day one and Marc Andreessen appeared on the cover of Time magazine, barefoot and grinning.

And then it all fell apart for reasons both internal and external, but many of which can be summed up in one word: Microsoft. Montulli cashed out in 1998 and went on to work at a couple of startups emerging from the Netscape talent pool: Epinions and Shutterfly. He spent a couple of years in Tahoe living the dream, “skiing and mountain biking every day” before getting lured back into the Bay.

Today he’s co-founder of JetInsight, a management system for charter fleets.

WHAT TECHNOLOGIES ARE YOU MOST EXCITED ABOUT NOW?

On-demand services are totally transforming service economies. Everything is getting swallowed up [by them]. Apps make things more efficient, and existing companies are slow to adapt to change.

Blockchain ought to revolutionize anything that deals with escrow services. A lot of Wall Street is just very sophisticated escrow services, so blockchain should have a big effect there. It will be slow because it’s such a regulated industry. But I don’t expect to make any money on bitcoin. [laughs]

WHAT ARE YOU MOST WORRIED ABOUT?

Well, I’d say the moral implications of automation on society. We might be approaching a tipping point where automation permanently replaces jobs. We’d be foolish as a society not to prepare for the impact, maybe through some universal basic income tests, creating a society not entirely based on your worth as a worker.

And I’m aware of the impact of the ad economy on media …. I’m paying for newspapers. I hope they survive.

Back at the University of Kansas, Montulli’s signature block contained a quote from Machiavelli. It captures the pioneer mind.

“For how we live is so different from how we ought to live that he who studies what ought to be done rather than what is done will learn the way to his downfall rather than to his preservation.”