Addressing the confusion around Windows 7 'build forking'

So there's been some silliness about the Windows 7 "build forking" stuff out in the blogosphere, which is to be expected. No one outside of Microsoft really understands how Microsoft creates Windows, but everyone seems to have an opinion about it for some reason, and some people even try to speak from a position of authority about how it's done. (Which I find odd, but whatever.) I tried to downplay this stuff on this week's Windows Weekly podcast because I don't think the process is necessary relevant from a conversational standpoint, whereas how Microsoft is changing the Windows experience is extremely relevant. But I had to address this issue in my Windows 7 build 7106 screenshot gallery because, obviously, it's confusing that builds are now jumping up and down between the 70xx and 71xx series. As a follow-up, I heard from a Microsoft employee who's emailed me from time to time off the record about Windows 7. And I'd like to present an edited version of what he told me here, if only to address the confusion out there. What it boils down to is that Microsoft has been forking the build since Windows 7 development began. This "forking" procedure is a bit different than before. The build number jump now is actually identical to the process followed for the Beta. For the Beta, builds 6980-7000 (or so) were reserved for potential beta clients, while development went on ahead for build 7022 as the first post-beta pre-RC build at the same time. It's not so much a "per-feature" issue Microsoft is avoiding here, working on individual features and slamming them all together in the end, but rather working towards a more stable public release. Once into Beta "Lockdown", and then into Beta "Escrow", the bar for fixes gets extremely high. Only the most severe, important fixes are taken, fixes that have the least risk to the stability of the Beta.  Any riskier fixes are postponed into the post-beta builds (7022+) to help keep the Beta as stable as possible. Now,