It’s masochism fortnight at Redmond!
Microsoft’s PR campaign of self-flagellation continues – with senior executives offering theWall Street Journal‘s Rob Guth an account of why Windows Vista will arrive so late and so incomplete. Thanks to the co-operation of Amitabh Srivastava, Brian Valentine, product manager Jim Allchin and even Gates himself – the Longhorn death march is turned into – wouldn’t you know? – a triumph.
The story’s subheads hint at both redemption – “Delay in New Windows Version Drove Giant to Develop Simpler, Flexible Product” – and punishment of the guilty: “Engineers Get Trip to ‘Bug Jail’”. However as we’ll discuss in a moment, there is no easy get-out-of jail-free card for the real culprits for the Longhorn saga: the senior executives who presided over Microsoft’s chaotic corporate processes for so long.
Redmond knows a thing or two about successful PR, and clearly wants to steer the press to the bad news in one lump. Business Week and Forbes lined up to chew the company out for late delivery of software last week. But this way, Redmond’s spin doctors hope, everything from now on will be part of a Microsoft comeback story. It’s a canny strategy that plays the home press corps sense of even-handedness … and short memories.
Which bring us back to the Journal, then. Last year Guth faithfully reported Bill Gates annual “reading week” while keeping a straight face throughout, and he’s come in useful once again here, in what might become an annual “masochism fortnight”.
Guth reports that Microsoft ditched the dysfunctional Longhorn, now Vista, code base in the summer of last year, reverting to a “clean” (these things are relative, we guess) internal fork of the Windows server code. This meant that WinFS wouldn’t ship with Longhorn, becoming the most important of a long list of features that had rolled overboard. But Srivastava’s teams introduced new testing tools to produce code more quickly, putting the project back on track. And lo, a Vista beta was delivered at the end of July with far fewer bugs than Microsoft expected, its executives report.
Gates and Allchin are even permitted to pat each other on the back:
“It’s such a huge step up from where we were,” says Allchin. “It’s amazing the invention those guys have brought forward,” adds Chairman Bill. “I wish we’d done it earlier.”
While Guth himself gushes, “The cultural shift is in swing”.
So all’s well that ends well. Or is it?
The Buck Stops … look! Over there!
The innocent reader, unaware of the Longhorn’s troubled history, might be tempted to ask why given less than a year and a “clean” code base, Microsoft couldn’t have shipped the Vista beta in 2002, or early 2003?
When Microsoft mapped out what Windows needed back in 2001, it did a pretty good job. Chairman Bill later PowerPointed these into “ten key Longhorn scenarios”, and the list included: Managed Code for better process security; WinFX to replace the ancient and crufty Win32 APIs; a better networked storage mechanism in the shape of WinFS; a graphics compositor to take advantage of the power of modern 3D graphics cards. Some were cosmetic, it’s true, and wouldn’t immediately solve the problems we outlined heremore than four years ago: “Why are my document files stored one way, my contacts another way, and my e-mail and instant-messaging buddy list still another, and why aren’t they related to my calendar or to one another and easy to search en masse?”
But at least the core Longhorn technologies would provide the foundation for addressing these issues in the future, and give Microsoft new competitive cards to play. Remember how it once planned to bundle SQL Server in every client? That really gave Larry Ellison something to worry about. Instead, the last few years have seen Redmond playing defense, not offense, against Linux, Salesforce.com and Google.
But at Microsoft’s Professional Developers’ Conference last week we learned that Microsoft is now evangelizing “gadgets” – and reintroducing a feature seen in 2003 Longhorn demos – the Dashboard! Given that the serious meat is not now going to ship in Vista, it’s eye candy and new device drivers all the way. And that’s very little to show for years of heightened expectations.
The problem now is that Microsoft develops software in its own image. At one time, when it was still in the shadow of IBM, the company saw itself as responsive, smart and user-focussed. Now that can take its customers for granted, the software Microsoft produces is as bloated and disorganized as the company itself. Windows XP ships with an OLE Calendar control most of the developers don’t even know is there. And six years after Microsoft shipped its scripting run-time – it finally got round to documenting it.
There are very few companies left who produce industrial strength commercial systems software: Apple, Symbian and Microsoft. And while Redmond’s rivals have had their share of problems over the years, few can dispute that their internal development processes are far more productive than Microsoft’s. Apple amortized Mac OS X development at $15 million a year: possibly less than what Microsoft pays people to run up and down its corridors with video cameras.
Few commenters responding to Microsoft’s star blogger agree that the week’s reorganization helps flatten the organization, or increase responsibility at the top.
“Windows 2000 was at least as big a mess as Vista is now. XP was basically the remainder of the 2000 feature list and it still took forever to get out,” writes one commenter. That isn’t an opinion – it’s a fact. Before Microsoft purchased a static code analysis tool developed by Intrinsia, Windows 2000 had over 200,000 bugs and looked dead on arrival. The WSJwould have us believe that new tools have once again saved the day.
But the inability to ship features as promised long predates Vista; the holes in the feature list simply remind us how many thousands of men years poor management has cost the company.
To lose one key Longhorn technology is unfortunate – but to lose all but one suggests that little short of a break-up, or fresh management from the top down, can cure these process issues.
Bring on the eye candy.