I just found a Vista Death March article by John C. Dvorak talking about Vista’s lack of ability to sell, and for the most part I agree with him.  However at one point he suggests they should go back and rebuild it on XP.  Funnily enough, that’s actually the exact reason Vista was delayed so much.  What I’m about to say won’t be a surprise to my close friends because I’ve been saying it to them for most of this year, but I think it’s time to try getting the word out there and at the very least give me an easy link to point people at.

The huge delay in Vista’s release cycle, and the dropping of all those cool features, did not happen because Microsoft are stupid.  It happened because Microsoft have more balls than most software developers (although money helps too).

The details are buried in Paul Thurrott’s Road to Gold article from last year.  It’s actually quite an interesting read if you have some spare time, but today we’re going to concentrate on page 3.  It talks about the 2004 area of the timeline, the point where Microsoft had been building hype about the OS for nearly three years and then suddenly went silent.  The story is spread a bit thin through the page, but is mostly summed up in this quote about a third of the way down:

By that time [April 2004], Microsoft group vice president Jim Allchin had decided that Longhorn wasn’t going to work. He told Bill Gates that the company would have to start over again from scratch, using the more recent Windows Server 2003 (rather than XP) code base as a starting point.

It was reported at the time as the “Longhorn Reset“, and while I remember reading articles about Microsoft’s new development practices I don’t think the implications of what had just happened really hit home for anyone (as evidenced by John’s article I linked above).  Microsoft had built Longhorn on top of XP.  They’d probably finished most of the features.  But the XP codebase was so old and fragile that the resulting system was horribly unstable.  So they threw out three years of work, began again with the Windows Server 2003 codebase, and took another three years to finish.

I don’t think anyone other than Netscape has had the balls to do that, which I’ll get to in a minute.

Vista definitely has a few issues, and it will probably end up with below average sales even though I think it’s fine and can’t stand the old XP interface anymore.  But Microsoft have deep enough pockets to afford this, and I think it had to be done.  In my experience Vista is far more stable than XP, much like Windows 2000 was more stable than Windows 98.  Hopefully a few years down the track we’ll have the vastly improved edition built on the rock solid base that the masses will love.

There is actually a precedent for my faith in Vista’s rewrite.  The classic example is Netscape who in 1998 threw out their code and created Mozilla.  Just about anyone who ever used it will tell you Mozilla 1.0 sucked (as did Netscape 6 and 7 which were basically just rebadged Mozilla releases).  However by 2004 that decision had produced FireFox and we all know how that turned out.

I’m not advocating everyone throw out their code; it killed Netscape and it’s going to cost Microsoft more money than they’d care to admit.  But if you have the resources to do it, I think it works.

Final thought – when describing Netscape 6, Joel Spolsky said the rewrite decision cost Netscape 3 years.  Is it ironic that that’s also how much time it cost Microsoft? 😉