That's an awfully subjective position to take. I don't have a problem with rapid release, but I have a problem with passing off minor revisions and bugfixes as major releases.

And try to tell me the version doesn't matter when there's a regression or otherwise something broken on a newer rev that works fine on an older one.

I don't know where this recent trend of "version numbers are evil" came from, but it's entirely silly.

Is it really silly? Gmail doesn't have version numbers. Facebook doesn't have version numbers. Hacker News doesn't have version numbers. Twitter doesn't have version numbers. They may refer to "the new version" or "the new, beta version" but not by number.

On the dev side, you might refer to a particular code snapshot with a "version number" that might as well be 3.2 or r304 or "2011-08-3 12:03:47" or d1b119d8f117b16fcfa58b2be60df87a6c45ac58. Considering the nightmares caused by conservative IT groups enforcing version X of software Y for years, eliminating version numbers from the public for something as critical as a browser seems like a good move to me on security alone. For other things the formal major.minor.release-revision can still make sense. Personally I miss the "even minor version represents stable, odd unstable" pattern that's increasingly fallen out of use.

>Is it really silly? Gmail doesn't have version numbers. Facebook doesn't have version numbers. Hacker News doesn't have version numbers. Twitter doesn't have version numbers.

Web apps and installed programs are completely different animals.