Mozilla, tortoise of freedom

After years in development, the Mozilla open source browser is nearly ready for release

After years in development, the Mozilla open source browser is nearly ready for release

In a way, this is a historic column, for it is still possible for me to say that at the time of writing the open source browser Mozilla has not been launched. But only just: the current version number of the Mozilla code is 0.9.9, and the first official version of Mozilla will see the light of day soon.

Once Mozilla does appear, it will lose an iconic significance that it has possessed for all of its already long life. Mozilla was important initially because it marked a huge victory for the open source approach.

When Netscape announced in January 1998 that it was giving away its crown jewels - the code to its Netscape Communicator browser suite - it was rightly seen as dramatic proof that free software had arrived. Significantly, it was that same year that IBM and Oracle also announced their support for open source - the Apache Web server in the case of IBM, and the GNU/Linux operating system for Oracle.

But if to begin with Mozilla was a shining example of how proprietary software companies might see the light, it soon turned into a textbook example of why going open source was hard.

As Netscape's offering was examined in detail, it became clear that it was a terrible mess. Not only had the program been bodged together to create the worst kind of spaghetti code, but there were gaping holes that needing filling. These had been left by major proprietary modules that had been ripped out when their owners had refused to allow their code to be released as open source. As a result, the Mozilla project had to do a lot of undoing and repairing before it could advance.

At this point Mozilla turned into a high-profile battleground as two schools of thought emerged regarding the way forward. One wanted to get the code into usable form as quickly as possible, even if it was still a little rough at the edges, the other wanted to ensure that the code supported all the relevant Web standards - something that would take much longer. In the end, the purists triumphed - not least because supporting open standards lies at the heart of open source and its success.

Although Mozilla has become something of a whipping-boy for those who are sceptical about the open source process, it is often overlooked that its slow and frequently painful evolution offers another important lesson - other than that free software takes time. This is that open source may proceed slowly, but it generally keeps going, and as a result keeps getting better.

Time will tell whether or not Mozilla manages to wrest back some of the browser market that its progenitor Netscape Navigator could once claim as its birthright. But the fact that Mozilla has managed to weather so many vicissitudes to meet its own high standards is significant, and a real indication of one of the key strengths of a development process that is not driven by commercial imperatives.

In fact, nearly all of the major open source projects have preferred to get code right rather than simply getting it out quickly. The timescale for the point upgrade between version 2.2 and 2.4 of the Linux kernel was two years, and version 2.0 of the Apache program has only recently appeared after years of gestation.

Aside from the general lesson that free software is more of a tortoise than a hare, there is another important implication of this pattern. The Openoffice project is based on proprietary code from the Staroffice suite, generously released as open source by Sun. In many respects, the Staroffice code is like that of Netscape Communicator.

It therefore seems clear that the Openoffice project - not to be confused with Sun's non-free Starsuite 6.0 which should be appearing soon - will proceed at a sedate pace. But, like all major open source projects, once it is finished it will probably be worth the wait.

Next week: Security

Read more on Operating systems software