Mozilla's Brendan Eich on the Birth of Firefox

A couple of weeks ago I posted the first part of an interview with Brendan Eich, who is Mozilla's CTO. That covered the early years of browsers at Netscape, and the origin of Mozilla. Somewhat belatedly, here's the second part of that...


A couple of weeks ago I posted the first part of an interview with Brendan Eich, who is Mozilla's CTO. That covered the early years of browsers at Netscape, and the origin of Mozilla. Somewhat belatedly, here's the second part of that interview, which picks up the story at the beginning of this millennium, and reveals the complex sequence of events that led to the creation of Mozilla Firefox.

One of the key people in this tale is Dave Hyatt, the main architect of tabbed browsing at Mozilla, and now at Apple. Eich explains: "he was getting fed up with Netscape management." Perhaps as a result of that frustration, he was also writing new browsers, one of which became the popular Mac OS X browser Camino.

"Dave quit Netscape to go to Apple," Eich recalls. "This was a real feather in Apple's cap – Hyatt knew all about Web compatibility – the team at Apple was very talented, but they didn't know about Web compatibility. He wasn't working on Mozilla at that point, except in his spare time. Inside Apple he was working on a fork of KHTML which led to the whole Webkit story."

Then someone else entered the scene. "Somebody named Blake Ross who'd interned as a high school student was at Netscape and he and Hyatt got along," Eich explains. "They were working on something called Mozilla/browser. The idea was to do just the browser - get rid of this 90s suite/Swiss army knife of browser, email, editor, newsgroup, address book" - which was what Mozilla ended up shipping after cleaning up the Netscape code. "They wanted to make something very lean and fast, because Netscape had larded the [Netscape browser] program with such fat it was slow to start."

With evident symbolism, the new browser project was named "Phoenix." As Eich recalls: "Phoenix 0.1 came out in the fall [of 2002]." At this point, Mozilla had just shipped 1.0 of its Internet suite. Eich says: "we finally reached enough quality that we were ready for widespread use – this was my tit for tat to Netscape 6 being rushed out two years earlier and not in good shape."

"Blake worked on [the new browser project] with Hyatt mentoring him, even though Hyatt was an Apple employee" - and having a few run-ins with his new employer. "Hyatt was blogging on his own blog about tabbed browsing and the sort of synthesis he'd done based on other ideas and older browsers, and at some point Apple told him 'we own that', and Hyatt said, 'no, there's lots of prior art'. They came to some understanding." Eich emphasises that "it was helpful to have Dave at Apple still put some energy into being a senior architect type to Blake who was doing the work."

However, "after Phoenix 0.4 or 0.5, Blake got really busy as a student at Stanford and stopped working on it. This was in 2003, and Mitchell [Baker] and I knew that something was going to happen to Netscape that wasn't going to be good. AOL had lost patience, AOL and gotten nothing out of the acquisition [of Netscape]. Time Warner got nothing out of that either." This was after the $164 billion AOL-Time Warner merger, which turned out to be a complete disaster.

Eich recalls: "we were sitting there in 2003 thinking AOL is going to lay off almost everyone at Mozilla, so what happens to Mozilla? The good news was that we'd grown the community so there was a better than Pareto 80:20 split: we had maybe 40%, Netscape 60%. Part of this was levelling the playing field - Netscape's new hires that weren't that good didn't get commit rights right away" - a shift from the previous situation.

Fortunately, the Mozilla team had some advance notice of what was going on in AOL. "Mitchell had gone to work for Mitch Kapor," Eich explains. "Mitch Kapor was friends with Ted Leonsis, an early AOL guy. They ran into each other at the very first D conference. Ted said, 'hi Mitch, I have this thing Mozilla, I don't know what to do with it.' Mitch said: 'I'll tell you what to do with it...' because Mitchell was working for him at the Open Source Applications Foundation."

"So that led to the creation of the Mozilla Foundation," Eich explains. They choose a foundation "because we want to be public benefit, we want to take donations, we want to be free of commercial agendas and the silly make the quarter numbers distortions that public companies face."

Fortunately, "AOL did the right thing," Eich notes. "It took some arm twisting by Mitch, because they were originally going to underfund us. Mitch said: 'no, you're not going to put them a lifeboat ethic situation where they make crazy desperate moves in the first year; you're going to give them two million dollars in two payments over two years.' So Mozilla Foundation was set up in the summer [of 2003]. The [AOL] lay-off came in summer. It was a big lay-off – Jonestown style. Everybody on a room, some hatchet-man executive saying: 'you're all fine people, bye'."

As a result of this mass redundancy, Eich says, "we had to think who were we going to pull out for the Foundation - a million a year is not going to pay for a lot of heads. So we picked some people who were not laid off at that point; there were given three months and they transitioned to the Foundation."

"We had four plus myself on technical side," Eich says – including Ben Goodger, now at Google. He had taken over from Blake on Phoenix, which had been renamed Firebird because of trademark issues (of course, it changed its name again because of unhappiness from the Firebird database project). There was also Mitchell Baker, Chris Hofmann, Asa Dotzler and a few others – ten or so in total.

But the new Mozilla Foundation also managed to keep in touch with other Netscape employees who had gone to work elsewhere. "The interesting effect right away was the community went from 60% being Netscape employed, to tiny." The people volunteering help came from a surprisingly wide range of backgrounds: "employees from IBM, Sun and Red Hat and I think also Novell and Oracle," Eich recalls.

As a result, "We had a pretty balanced community - we had a respectable open source project . It was no longer dominated by Netscape," he says. "It had a level playing field, it had evolved governance rules, so that people knew how to behave and how to get access and how to revoke it if you needed to, which really doesn't happen that often. They knew how to appeal decisions that were delegated to module owners up to me or governance issues up to Mitchell. It was pretty healthy. The only issue was: will we survive?" Money and resources were limited, and needed to be husbanded carefully.

"So we started focussing on Firefox," Eich says. "Because even back in 2002 when Phoenix was born, Mitchell and I met with Blake and Hyatt – he was at Apple but he came over to Netscape one day – we said: 'we 've Mozilla 1.0, we've done as well as we can with this suite, this Swiss army knife; we'd like to focus on what you guys are doing.' And Dave Hyatt and I wrote a Mozilla roadmap in 2003 April that said: let's do just focussed apps – the browser, thunderbird the mail user agent – let's build them with minimal UI incrustations, and instead have an add-ons architecture."

Eich explains the reasoning behind that. "In the open source world you have a lot of people who want to scratch an itch; in fact they often want to scratch the same itch over and over. They will actually enjoy rewriting the code base to make an incremental improvement or even aesthetic change – they are not adding user value. And other people will want to add user value for themselves but it's very peculiar. We're not going to put that into the default user interface of something that has millions of users – even Mozilla had millions of users – so you simply cannot keep everybody happy. And you end up with this contested tragedy of the commons problem."

"The only way out is to create an extension mechanism that is easy to author and has no barrier to entry and adoption, and has a sort of intentional supportable model for having hundreds of add-ons - or thousands, as we have. Add-ons are necessary to reach the mass market and to cross the chasm." Moreover, as he notes: "in fact, all the browsers have add-ons now." It was one of the first times that Mozilla led the way – but by no means the last.

Follow me @glynmoody on Twitter or, and on Google+