Fair enough. I would say, though, that the campaign would probably never have worked if Firefox was only just as good as the bundled IE. User inertia keeps the bundled product dominant. This must be true since Firefox still has but 21% yet is arguably superior to IE in nearly every conceivable way. The point I was getting at was that bundling can have pronounced market-share effects.

I do think that your point regarding "no bundling at all" as a philosophy is extremely valid; however, I wonder if it is pragmatic or realistic. I think it is unavoidable that a browser gets bundled because user expectation is that they will get the "internet experience" out-of-the-box. That sentiment applies not just for Windows, of course, but for every major desktop and distro.

Which leads to a contrarian question: under the article's logic, should major FOSS desktops be bundling Firefox?

Finally, I am not sure that the claim that bundling is bad for competition is tenable as a general rule. How can Firefox exist today in its current form if that is true? I wonder if the answer lies in the fact that bundling, while bad for closed-source entrants is not particularly a deterrent for innovation and popularity when it comes to FOSS projects. In that light, I can better understand why Mozilla may not have an interest in bundling but Opera does.

You mean bundling IE did not kill Netscape, which at the time, was the dominant browser? Yes, it assuredly did. Bundling provably leads to market share when the bundled product has equivalent or near equivalent properties of the alternatives. IE currently lags far behind the alternatives -- so of course there is room for competitors. Still IE manages over %60 share and its peak (prior to *compelling* alternatives) had over 90% share. This can only be satisfactorily explained by the fact that IE was bundled with the OS that was bundled with the PCs that users were buying.