Why So Little Hoopla Over Windows 7? It's Partly By Design, Says Windows Engineering Chief Steve Sinofsky
Posted by: Peter Burrows on October 30, 2008
So where’s all the fuss that usually accompanies the announcement of a new major release of Windows? The raging debate about each every last feature? The endless analysis in the press? Well, there’s some obvious reasons. Somewhat more pressing matters like the election and the economy are keeping Windows 7, which was previewed by Windows engineering chief Steven Sinofsky for the first time at Microsoft’s Professional Developers Conference on Oct. 28, off the front pages. Also, Windows 7 was overshadowed (at least to me) by the unveiling of Azure, the “web-tier operating system”—which will define Microsoft’s move into cloud computing. Lastly, as my colleague Steve Wildstrom points out, Windows 7 just isn’t all that exciting. It seems to be more about fixing the ills of Vista and improving the experience of using existing capabilities, than taking any giant leaps forward. (Judging from comments from developers as I trolled the halls of PDC after the demo, this is just fine with them. “It’s just an operating system. I’m glad they’re paying attention to the fundamentals like performance and reliability, rather than marketing,” said Jan Oudkerk with Veritmart BV, a Dutch company that makes software for dentists. And of the new features, Ottawa-based consultant Bruce Johnson was most impressed with the fact that his work laptop will automatically connect with his home printer when he walks through the door. “Now I won’t have to worry about printing my resume on the office printer any more.”)
But there’s another reason for the relative calm, says Sinofsky: a radically different, more carefully-controlled process for developing (and communicating with the external world) about Windows. Here’s a short Q&A from an interview I did with Sinofsky the evening before his keynote (my questions have been edited for brevity). Read on, after the break.
Q: So there don’t seem to have been many leaks regarding the finer details of Windows 7.
Sinofsky: Surprisingly, no.
Q: Why is that? Did you do something different this time around?
Sinofsky: It’s one of the things we learned from the Vista product cycle. [During that process] we were out there talking about things before we quite had the design worked out. There are times you need that [approach]. But the feedback you get is based on ideas that haven’t been fully formed, even by us. That’s very hard to reconcile, because people are projecting based on where they think things will finish. Show it to 1,000 people, and that’s 500 places people think it might end up. You can’t incorporate that feedback in a systematic way, when you don’t even quite know where you’re going. Second, people are never quite sure what you’re going to end up with. Will this feature or that feature be in or not be in. We worked hard to get to the point with Windows 7 that the designs are fully formed. So when we show it to people, we feel confident that we’ve thought through many of the issues people might run into.
Q: Do you think waiting until you had defined your priorities to share details broadly helped you get to a better result?
Sinofsky: It’s true for any engineering endeavor. If you’re remodeling your kitchen and your architect tells you all about all the things you can do even before you’ve set a budget or set a baseline, you can get very excited about that 30 x 30 island that won’t fit in your kitchen.
Q: So how was the process different?
Sinofsky: In the past, we’ve often given [information] to everyone very, very early. So there was this broad conversation [in the marketplace] about what’s in, what’s not in. It’s been happening for years. Even the PC makers have often been confused…We didn’t have a systematic relationship with the OEMs [where there feedback was incorporated into the process]. We had a very high bandwidth conversation, but there wasn’t a systematic feedback and planning process. With Windows 7, we sat down a year and a half ago and said, "here’s what we think is a good development process. We’re going to establish a common baseline before there’s any code, and then every quarter we’re going to get together for a week to keep checkpointing where we are." In the [software development] world, we talk about a design funnel—where you have a huge opening and a lot of ideas pour in. But you have finite resources, and finite time, and you need to come out with a coherent product. This time, we brought the PC manufacturers through the funnel with us, so we’re all partners together.
Q: I’m interested in hearing more about how you managed to prevent leaks. Did you have any special NDAs, to keep people from talking?
A: The goal was not to keep it secret, per se. It’s about mutual respect. We are working on an early stage product, and we don’t want people to see our half-baked work. We’re a bunch of chefs collaborating on a great dinner, and having a bunch of people run through the kitchen halfway through is not necessarily the best way to do a great job.
Q: So how many PC makers attended those quarterly meetings?
Sinofsky: There’s a finite number of PC makers, but still, we can’t work with all of them. But we talked to all of the leading PC companies.
Q: How many people attended the meetings?
A: We think of this as engineering to engineering. At every PC maker, there’s people whose job is to come to work and make PCs. I know that’s hard to believe [given the huge amount of outsourcing], but it’s true. So depending on the company, they’d send five or ten or twenty people to Redmond…We had different kinds of relationships with other types of partners. The graphics people, for example, have people full-time [at Microsoft’s campus). They need to be in our hallways, because it’s super sophisticated. Building a graphics card is rocket science.
Q: So far, given what you’ve told me about Windows 7, it sounds like the major appeal to consumers will be to fix some problems with Vista and improve the experience of using current features, such as home networking.
Sinofsky: This is a conference for developers, and this is just a pre-beta. But after WinHEC (The Windows Hardware Engineering Conference, which takes place in San Francisco in early November), we’ll start to amp up the [messaging about the] broad consumer appeal.