News

Exit Interview: Sam Ramji Assesses Microsoft's Open Source Ascent

During Ramji's tenure at Microsoft, company instituted interoperability principals and improved ties to Linux, PHP communities

• By Jeffrey Schwartz
• 10/01/2009

Sam Ramji's five-year tenure at Microsoft was marked by, among other things, the company instituting interoperability principals and more recently gaining closer ties to the Linux and PHP communities.

Shortly after Labor Day, Ramji announced that he planned to depart Redmond to return to Sillicon Valley, where he would join cloud infrastructure vendor Sonoa Systems, as reported here. In his final three months at Microsoft, Ramji said he achieved his long-held goal of launching a non-profit open source foundation sponsored by Microsoft. The CodePlex Foundation was launched in conjunction with the announcement of his departure. Ramji will serve as the foundation's interim CEO for its first 100 days.

Can you describe your last few weeks at Microsoft since your move and the foundation were announced?

I spent most of the last three weeks launching the foundation. I had very little time for anything else. I was dealing with very mundane logistics like getting a tax ID, and opening a bank account and all the things required of creating a startup that you don't necessarily associate with a non-profit foundation. We had a ton of coverage in media, blogs, journals, Twitter, and our Google group. In general people are pretty optimistic about what the foundation can represent. There are always a few people that are going to be concerned no matter what the topic if it's something about Microsoft and open source, but that was a small minority. The majority was optimistic.

We had some people from both what I think of as the core open source community as well as specific communities like the Apache Software Community or the .NET Open Source Community, who immediately got engaged and sent recommendations and asked questions. We've gotten some very good advice on how to structure the foundation, on where to have our initial focus and we will continue to get out to the public and to ask questions. We will be at the MonoSpace Conference on October 29th talking about the range of .NET open source developers, and we'll be at the Apache Conference the week of November 2.

Speaking of advice that you've gotten, can you discuss your reaction to Andy Updegrove's advice and concerns, which generated a lot of discussion?

Andy made a number of salient points based on his early extensive experience with foundations for standards. I appreciated his feedback. We've discussed it and look forward to taking some of those steps such as becoming a membership-based organization. He had suggestions on expanding the size of the board of directors, we think that's important. There's a tremendous amount of open source being used in academic institutions, but also some of the challenges academic institutions have had in making contributions back to the open source community projects have echoed some of the challenges we see with software companies. So we think the foundation can serve as a useful conduit for them to be able to start to contribute back to some key projects.

Was there anything you take issue with or disagree with?

It's an education for me, frankly. Four of us including myself are new to foundations. The amount of time you need to spend analyzing the bylaws and formulating responses, we all saw was a very useful education. The one comment I had for Andy was we didn't plan to launch it in a perfect state. We launched it at alpha or beta, depending on your point of view. I think we are going to get more independent, different software companies and communities participating, especially sponsoring. I think he had a very good point about financial independence.

Whose idea was it to start the foundation and how much buy-in did you get from top levels at Microsoft?

Bill Staples is a key figure in the formation of the foundation, Mark Stone is our deputy director. There was a pretty good size team of people across legal, my team in the Platform Strategy Group and people in the developer division and the server and tools business. Of course, we had to plan executive sponsorship right up to the senior leadership team. Collectively there's an open source community within Microsoft for some years. We started having this idea in earnest probably a couple of years ago.

I named Sandy Khaund. She wrote a paper called Microsoft.org which was thinking about the great things we could do for Microsoft and for open source if we were to establish a non-profit, open source foundation. So we took a stab at starting a foundation a couple of times before, but weren't successful getting those projects off the ground. This time around we had really strong buy-in and strong leadership from product groups from people like Bill Staple that helped make it successful at this time.

How long did it take to pull it together?

We started running in early June and we sprinted through most of the summer to reach our launch date of September 10. It was a solid three months.

What happened this time that made this effort work, where past attempts failed?

I think the idea of open source being complimentary to Microsoft has just gotten so much broader acceptance. Additionally, the work that Bill Staples' team has been doing with the open source communities, like PHP applications, as they look at what are developers doing today, how are they building Web applications, what types of technologies and languages are based on teams. [They] built a real strong competency in understanding how open source is driving the Web application model. Everyone has gotten used to the idea that open source and Microsoft have a complimentary future and there is a real strategic push by an organization that focuses heavily on developers.

Did this reach up to Steve Ballmer?

Steve was briefed.

What was his initial reaction?

We had no resistance.

So can you say you have left Microsoft with the feeling that the company truly is grounded in this new open source mission? You didn't leave with any frustrations that things were happening too slowly?

No, in fact in the last few years, and last few months especially, the slide wheel has been going faster and faster, and I would say we've gotten some of our fast results quite recently. For example, when we released the Linux kernel drivers for HyperV under the GPL, we achieved the PHP 5.3.1, which is the fastest PHP on Windows in June, so there were too big hits in a short amount of time.

I am not saying it happened overnight -- it was pushing on four years or so -- but more and more people have gotten on board with it and it started to go faster and faster. What I expect to see from Microsoft in the coming 12 months would probably be conservatively twice as much open source contribution activity than we saw in the previous 12 months. The open source concept is really spreading across the engineering teams inside the company. It's really becoming part of the DNA.