The public face of the Windows development-chief duo is Senior Vice President Steven Sinofsky, who runs Windows and Windows Live Engineering. The private face is Jon DeVaan, the head of Microsoft's Core Operating System Division (COSD).
DeVaan and Sinofsky have worked together for a considerable about of their Microsoft careers. DeVaan has kept a very low profile. (So low, in fact, that his WinHEC keynote on November 5 was his first public speech about Microsoft-related matters since 2001.)
So who is this masked man? I had a chance to chat with DeVaan by phone on November 5 after he finished his remarks on Windows 7 at Microsoft's hardware-engineering conference.
I already knew from DeVaan's bio that he'd been with Microsoft for more than 20 years. He's held a variety of jobs at Microsoft -- everything from managing the TV Division, to co-managing the company's Consumer and Commerce Group (where MSN was housed, at one point). He's best known, like Sinofsky, for his work on Microsoft Office. He worked on Office 95, Office 97 and Office 2000. He also started and was in charge of the company's Engineering Excellence initiative. He became the head of COSD in 2006. (Among other functions, COSD is charged with developing and maintaining the code shared by Windows client and server.)
Here is a bit more about the silent partner behind the next versions of Windows. This is Part 1 of a transcript has been edited for length and clarity.
JD: So, engineering excellence is interesting in that I'll kind of refer back to one of my favorite places in my career, which is way back. You might remember it was in the days between Windows and OS/2 when Microsoft was really struggling to ship its software. I don't know if you remember how late Word for Windows was originally or any of those adventures that we had, but it was really a point in time where we had to take a step back as engineers and look at how we were doing things and create new ways of working. And we created a lot of the basic mechanisms that have served Microsoft incredibly well really through today.
When I was asked to take engineering excellence for the company, working for (Chairman) Bill Gates -- this was in, what, 2000 or 2001 -- we saw with Longhorn (Vista) that this was kind of a time when we needed to do another take like that of stepping back and looking back and reforming how we do things as a company on engineering.
You look at products like Office 2007, you look at things like the latest Visual Studio, so Visual Studio 2008, SQL Server 2008, those are products that we impacted by taking that step back and looking at ways that we need to improve and get better.
And now in working on Windows, my job is to take those ideas and instantiate them for Windows. MJF: Is Engineering Excellence is a superset of COSD?
JD: The right way to think about that is when you look at a company with the breadth of offerings like Microsoft, there is no such thing as an engineering system that's the right engineering system for all those cases. Exchange Server is not Xbox, and to approach the engineering on those products in exactly the same way wouldn't be right.
So, what you see with the corporate level with engineering excellence is kind of the keeper of best practices and the facilitator of sharing across all of those unique engineering systems at the company, so that we drive the best ideas to the places where they'll do the most good.
And so my role on Windows now is I am one of those engineering systems, and we're going to innovate in a way that makes sense for Windows to drive quality and agility and all those things we want to be able to accomplish. And the ideas that we generate in Windows will get rapidly shared to the right places across the company. MJF: Were there are any specific learnings from when you were with the Office team that you've brought with you to COSD and the Windows side of the house?
JD: I think there are, and I think there are also experiences which have been molded not just by Office, but by the larger set of good ideas from my work on engineering excellence.
When you look at some of the ideas that come from Office, I think the well defined engineering project with a clear vision is clearly one of the engineering excellence concepts that I feel proud -- actually Steven (Sinofsky) and I together pioneered in Office.
The way we organize the engineering teams to build long term depth of the engineering capability in the team is another one of those ideas.
Then there are ideas that come from other places, like the incubation part that I'm talking about really probably comes from the SQL team more than the Office team, and the idea is when you're in a situation of taking a step back and trying to learn from your experience as engineers, why not take the best ideas you can find period. Sometimes the situation is different and old ideas might not be the right ones, and you need new ideas. MJF: What do you consider your greatest achievement at Microsoft to date?
JD: It really has to do with something that I've mentioned a little bit already, and it's returning to that time when we were really struggling to ship software at Microsoft. So, the way that we thought about that and developed practices that if you told me then what we've been able to accomplish with those basic practices as a company, I probably wouldn't have believed you, the notion of the feature team and the daily build and milestones and all these basic engineering ideas that have served the company incredibly well.
The first time we employed them directly for me was on Excel 3 where, if I remember correctly, we were 10 days late on a two-year engineering plan, which was an amazing record for the company at the time.
So, I'm really proud of that, and that translates into the work that I've been doing on the company on engineering excellence, which is doing that same kind of thing at a broader scope.
Now, proof in the pudding time here working on Windows, and I think we're making a lot of progress, and I'm really encouraged by the kind of response we're getting from partners as they see the quality of the release, and the different tone in how we're working together. MJF: Given your consumer heritage at Microsoft, I'm curious whether you feel there is any merit in the idea of -- at some point down the road on Windows -- of having separate code bases for more consumer-based versions of Windows and more busines- based versions of Windows?
JD: I'm not a fan of that idea. Of course, it's impossible to say never. There are a couple basic reasons. One is we're always dealing with actual people, and actual people welcome consistency in their life. So, it's actually I don't think we're really serving people if their work life is a lot different than their family life or a consumer life. I think people really want some things to be consistent across there, and if we're able to do that effectively, we're helping people a lot more.
The other one is the expertise to do these things doesn't grow on trees, and so I think when you take an approach like that, you dilute your expertise and create problems that are not foreseen when you think it's time to make that kind of a decision.
The answer to this question may not be right for everybody, but I'm not a fan of it for Windows. We can deliver a lot of value with that consistency and integration.
(There's more to come. Part 2 of this interview will highlight more of DeVaan's views on Windows 7 and beyond.)