Microsoft's Tech Summit: Redmond Still Trying To 'Get' Open Source Software

It isn't often that I get to spend several days listening to the people that are creating the software products at Microsoft. In the past, when I have been to Redmond it was usually on a press junket with colleagues from mainstream IT and business media, listening to spin doctors and marcom folks who were trying to get "coverage."

Last week, I was in the company of people who write code, and have a deep knowledge of the tools, software, and applications, as part of the company's Technology Summit. I wanted to jot down some thoughts, what I learned, and what the role of Microsoft will be in the coming years for the evolution of the Web and open source.

Microsoft still has this love/hate relationship with the Mac. Some of the presenters deliberately brought Macs, including one who ran Vista under Parallels. Yet when someone asked if anyone had tested interoperability of Vista with Mac OS X, it was clear that this wasn't a focus. (And for those of us that have both, a continuing frustration.)

For Microsoft to succeed, the Mac has to move from being a poster child operating system, a necessary evil, and some annoying relative to be tolerated to an actual strategic direction and integral to the company's success. The people that have moved to Mac desktops are canaries in the coal mine. They aren't happy with Windows for very real reasons (blue screen and security sinkholes come to mind). It continues to be a platform that is used by many developers.

There is a growing emphasis on interoperability at Microsoft, and they are clearly spending a lot of resources on projects (such as Windows and other operating systems, new versions of Windows networking protocols, and new programming languages with older ones), but there is still room for improvement. You can never do too much interop testing. Interop is getting more attention, but still isn't infused into the core culture yet.

Microsoft is a company of coders, and they respond best to an audience full of coders, too. Coders are the heart and soul of what drives this place. They have always understood what developers do and think and eat and drink. Speaking of which, during one of the presentations, an Outlook reminder popped up on screen that listed as overdue the items "eat dinner" and "go home." That resonated amongst the geeks in attendance.

But let's face it -- in the past several years, developers have moved away from writing code for single-PC applications and Microsoft still doesn't quite get this whole Internet thing. "We didn't understand open source and didn't use the correct words back in 2005," said Bill Hilf, one of their head open source advocates.

During the meetings, the audience took them to task about lack of enthusiasm for various open source projects. I found it interesting that most -- not all, but most -- of the presenters still were viewing open source as competing with some Microsoft product offering. They need to realize that people are going to use both, and want not only choice but also the ability to freely code in both Microsoft and open source projects.

Don Box, one of the developer evangelists, semi-seriously said, "I humbly apologize on behalf of the 70,000 owner-operators of Microsoft for the statements our CEO makes to scare all the open source people." But there was an element of truth behind it. The more that Microsoft can make this ambidexterity possible and successful, the more software it will sell. Some of the presenters clearly understood this, but others still characterize things as "us" versus "them."

Because Microsoft is a big company, it is hard sometimes to identify when or how a particular program or project will ultimately drive bottom-line revenue. Is getting more people to write .Net code going to bring in more bucks than getting more people to buy more Windows servers? Is getting Windows better at running PHP going to drive more revenue than getting Windows to become a more secure Internet-facing operating system? I dare say that these aren't easy or simple decisions, and sometimes they don't get it right the first time.

What is clear is that Microsoft "is the most fanatically self-critical company that I have ever worked for," said Hilf. They spend time even examining other people's code, just so they can learn from their mistakes, at least according to Michael Howard, their security czar.

"But I don't want you to love me, I just want you to buy more of my software, " said Hilf. Note in that statement is the assumption that we are already buying their stuff. But not enough: "We have failed to convey the power of our platform with the elite," said Sanjay Parthasarathy, the über-evangelist and programming manager.

They are trying to regain Web thought leadership with IE7 and IIS7, but the open source group (or at least the group that was assembled) has moved on to Firefox and LAMP. "Seventy percent of the Web sites are scripted with PHP and under 20% of those are deployed on IIS," said Sam Ramji, the director of their open source labs. "We are losing these developers and doing something wrong."

Many of the attendees that I spoke to had a "nothing to see, let's move along" attitude about IE and IIS: they haven't used the new versions, didn't really care, and weren't interested. I surprisingly learned that a full copy of IIS7 has been shipping in Vista -- did I miss that memo? Gotta wonder with all the stuff that I read (and wrote) about Vista, why this key factoid eluded me until now.

Part of the problem (for Microsoft) with Web development today is that it is too pluralistic. Microsoft thrives best when it can focus on a single competitor -- Don Box mentioned how they are laser-focused on Google, made even more ironic by a developer who works for Google sitting right next to me.

"We are a lot of little companies inside here and one of them will figure out a way to crush Google. Still, they are the best thing that happened to us, and are going to make us better." But focusing on Google isn't the only answer, and the problem with open source is that a thousand flowers are growing out there, and maybe 10 or a hundred of them will bloom and blossom into something useful. It is getting harder to keep track.

A corollary to this is that the circa-2007 world of programming is all about being able to teach new programmers how to learn new languages. This is somewhat of a challenge for the compsci departments of today, who are trying to find a new curriculum and state of purpose for their students.

No one knows this better than Microsoft. Kevin Schofield, who runs Microsoft Research, called Microsoft "the world's largest compsci department." They have published almost a paper a day for the past 15 years.

It was quite a learning experience this week. I apologize if these are more like notes than a coherent essay, but I am still digesting what I heard, and reading the various blogs of the attendees and presenters. I have posted links to all of these discussions (Ben Galbraith and Travis Swicegood have the most complete coverage of the MTS meetings) on my blog.