was the first example of a company releasing the source code of a very large program to the "Bazaar"
development model, while continuing to work on that code itself. Any company that is considering following Netscape's lead should pay a great deal of attention to what the Mozilla Project got right, and what it got wrong.
The initial code release was generally of poor quality, and was missing several key features that Netscape could not have included due to licensing issues (for example, there was no Mail/News component). After a year of wrestling with this codebase, it was decided that the only course of action was to throw away the dead code, and for almost all of the suite RewriteCodeFromScratch
. Many developers thought this was a stupid idea.
The big mistake was that while they changed their goal from "Upgrade a browser" to "Write a new browser", the project did not adjust their release scope, and did not adequately communicate with their "customer" (the public) about the implications of this. Mozilla had become a huge task.
The project was faced with a mountain of features that had to now be implemented from scratch, but the perception of the project was building an upgrade, rather than a new product. New browsers (for example Konqueror) can start off with a limited feature set and grow incrementally. Anything carrying the Netscape brand would have to have an improved feature-set over Netscape 4.x, and ideally would have to be competitive with the most recent version of Internet Explorer, which now had several years' head start. The result was years spent in the wilderness, with nobody really believing that anything worthwhile would emerge.
However, as of 2004, the MozillaBrowser
is the de facto standard browser on Linux, and its lightweight cousin, MozillaFirefox
, is now making inroads on both the Linux and Windows platforms.
Bazaar vs. OpenSource
I apologize in advance, but I'm going to do an annoying OpenSource
advocate thing and quibble over terminology. Having a project be OpenSource
does not automagically make it conform to the bazaar model. EricRaymond
's well-known critique was directed as much at the FreeSoftwareFoundation
as at proprietary software. Linus, he observed, took patches from a wide variety of people, while FSF hackers tended to be more exclusive. The point wasn't just slapping a license on a codebase, but building a real community around it.
- I don't think that's a nitpicky quibble; it's an important point.
Mozilla had a lot of trouble attracting outside programmers. To a lesser extent it still does. I don't know if it was a corporate-cultural problem or not; definitely the monolithic size of SeaMonkey?
The word 'bazaar' was chosen deliberately. EricRaymond was cited as a significant influence on the Netscape management when deciding to OpenSource the product, and the Mozilla project fulfilled all the requirements of the bazaar: open communication, openness to contribution from outsiders (many of whom were in fact later hired by the Mozilla project itself), and release early, release often (Mozilla published daily builds). The fact it had trouble attracting outside programmers was mostly because it was a huge problem, with a steep learning curve. There wasn't enough low-hanging fruit for new developers to pluck. What wasn't bazaar about it?
Many have complained that it's just built a bloated mess. It's huge. It's bloated. It's extremely late. It's a failure.
complained that ... but IMHO most of the detractors don't understand the scope of the project. It is
large, but not particularly bloated. The design is much
cleaner than anything else out there. It may be late by the original projections, but once the (correct, IMO) decision to drop the original codebase was made, you could hardly expect those projections to be made.
Always amazes me that commercial projects delivered amazingly late (i.e. Windows98, NT5.0 (nee W2K), Java2, J2EE, etc.) and amazingly bloated to much press adulation and fanfare without actually contributing very much to the commonwealth are lauded and a project like Mozilla (ignoring for now that its goals were overreaching) should attract so much negative attention. When all is said and done, the Mozilla project will provide a nice standards compatible browser that works on a multitude of platforms and a mass of source code and components that will be gleaned for years to come. What did you get with IE?
It's attracting so much negative attention because it was held up as THE MODEL of bazaar-style development and how it changed NetScape, blah, blah. The only thing it's done is put NetScape out of business. How many ways can you spell failure?
released Mozilla as OpenSource
it wasn't their main market anymore. They were more into servers by then, so I find it hard to believe that Mozilla put NetScape
out of business. -- MarkoSchulz
How exactly is NetScape
out of business? NetScape
was bought by AOL, and is still part of AOL, which is still very much in business. Sure, we might like to complain about the banal people in charge of AOL/TimeWarner?
. But I'm not sure how bought out by soulless megacorporation
== out of business
. -- FrancisHwang
Did they sue Microsoft because NetScape
lost the market to them (according to their beliefs because of unfair competition)?
I have to admit that 'market' is a very fuzzy term. I thought of it as 'making revenue', in which other things became more important to NetScape
than Mozilla IMHO. But I have to admit that if you look at it at an different angle, such as market position or power by control, the result can be quite different. -- MarkoSchulz
Once upon a time, NetscapeNavigator
was shareware. You could download it without paying, of course, and many people did, but Netscape still made more than half of its revenue from companies licensing the browser, shops selling it in boxed sets, and even the occasional person registering their downloaded copy. When Microsoft released their FreeAsInBeer
Internet Explorer, Netscape was forced to make Navigator similarly free.
Microsoft had enough money to support the continued development of a loss-making browser. Netscape didn't. In desperation, unable to afford to write Netscape 5 on their own, Netscape released the source in the hope that it could use the OpenSource
community as free developers. JamieZawinski
hit the nail on the head when he described why that didn't work, although now that Mozilla is looking more and more like a real program, it's finally seeing more non-Netscape developers, Although when I say non-Netscape, I mean non-AOL, since Netscape ran out of money, and is now the Netscape Division of AmericaOnline
. -- CharlesMiller
for a point by point analysis of the reported failures of the Mozilla project.
In a June 2004 ArsTechnica
interview, Scott Collins says a Netscape 5 was almost ready to release, but a senior executive (he doesn't give a name) nixed it and gave the development team three months to release a Gecko-based version. Perhaps had Netscape released a version 5 to compete with IE, the years-long gap between the old product and the Gecko-based product would not have been, and the browser wars would have taken a different road. Neither the decision nor the result can be changed, but knowing if there was indeed management blundering might be instructive.
Yes, that was an incredible blunder. It is only in the last little while that Mozilla has again had a superior (to IE) browser ... but without the market share to do much with it. Perhaps MozillaFirefox will get enough uptake on MicrosoftWindows to make a difference; who knows.