(Hat tip to Neowin)

Steve Ballmer is promising a more lively Windows release cycle, i.e. it won’t take five years for a new release to come out anymore. I remember hearing around the end of the Windows 2000 development cycle, which took a relatively short three years (by comparison) that they were going to stick to more incremental releases. Windows XP is a fantastic example of this strategy gone right: Within a year of Windows 2000′s release, they were able to unveil a well-tuned, faster, and more reliable update to Windows 2000 that has, in my mind, outlasted its lifecycle but still manages to work really well with the right patching and maintenance. Windows Vista is an entirely different story.

Is this a bad thing? I wouldn’t be so concerned about Windows Vista’s gestation time if half the features they had promised in this five-year development cycle were actually going to be delivered. It seems like for a while last year they were cutting a new Vista feature every week in order for it to make the January 2007 release Bill Gates is now hoping for (80% chance). If they were still developing the original Longhorn vision, then perhaps I could understand a five-year wait, as Longhorn was a rather revolutionary product. Despite a lot of its internal changes, Vista is still very much more of an evolutionary improvement from a user and developer standpoint. You can rearrange UI and make it pretty all day long and add new security blocks to make it impossible to rename a shortcut without three administrator approvals (I hear they’ve improved this now), but the hooks that were there for developers are largely diminished. There’s hardly any managed code in Vista, Avalon is unbelievably absent…it’s sad, really. Longhorn 4074′s unified notification system and sidebar would have been truly useful. In Vista, I’ll turn the sidebar off, as it does nothing for me now, and forget about that one source of notifications, which was even in Longhorn 4074 unbelievably elegant. Nope, Outlook will still have its own way of popping up new e-mails that come into my inbox, the Found New Hardware Wizard will still use a balloon to tell me that it couldn’t find drivers for my sound card, and Windows Live Messenger will still use its own sliding popup window in the corner. In other words, Windows Vista is still the total user interface clusterf*** that Windows XP and its predecessors are, despite the two years of time developing and promoting the resolution to this problem only for it to suddenly disappear in Build 5048.

Microsoft has a lot to learn from this, and I’m sure they will. Two things they should have done differently:

1. They should have, without question, started the project from the Windows Server codebase from the beginning, not doing a significant amount of work on top of Windows XP only for them to totally scrap it and literally start over. It would have been much easier to merge in the SP2 changes to the Longhorn tree versus the other way around. The “reset” was a giant, costly mistake that will ripple all the way to Blackcomb (if Microsoft doesn’t stick yet another project before it i.e. Longhorn and Vienna).
2. Microsoft needs to know when to contain its excitement over pre-alpha vaporware. It got developers and IT junkies alike excited, and the letdown with Build 5048 was pretty crushing, IMO. So yeah, time to get the developers more excited later into the development cycle, when a realistic final form of the software can take place.