The real reason behind Microsoft's leap into open source

Microsoft has created a new open source subsidiary. Is this a sign of greater openness, or are there darker motives?

At the end of last week, Microsoft quietly released unexpected news: the launch of a wholly owned subsidiary to engage open source projects. All the reports stuck to the facts the new president of Microsoft Open Technologies, Jean Paoli, was willing to disclose in a blog post. I approached Microsoft for more information -- especially on why the company took this step -- but received a "no comment" beyond Paoli's blog. Given the new company includes all of Microsoft's open source evangelists, none of whom are exactly shy, I found this reticence remarkable.

Microsoft's history with open source is checkered at best. Its initial strategy in the late '90s, leaked in the Halloween Documents, was to attempt to undermine and attack open source, partly by funding SCO's attacks on Linux. Back then Steve Ballmer famously called Linux "a cancer that attaches itself ... to everything it touches" (actually meaning all open source and not just Linux) and has never to my knowledge recanted.

[ Delve deeper into Simon Phipps' bold stance on open source in "Why software patents are evil." | Track the latest trends in open source with InfoWorld's Open Sources blog and Technology: Open Source newsletter. ]

But Microsoft's overt hostility has given way to a more pragmatic approach, at least on the surface.

Microsoft has come to the realization that open source is an inevitable part of the marketplace and has instead tried to triage it, first at arm's length, then increasingly through open source projects. Indeed, Microsoft is the 17th largest contributor to Linux, hosts project at its nonprofit foundation, supports the Apache Software Foundation, and regularly shows up as a sponsor of open source events.

Open source contact points are now all over the company, though no core products truly adopt an open source approach. Behind the scenes, however, Microsoft continues to subtly undermine open source, as demonstrated by this week's FOIA-backed revelations from Glyn Moody about how Microsoft lobbied against open standards in the United Kingdom.

The new subsidiary is another evolutionary step in Microsoft's open source pragmatism. Since, as Paoli is careful to say, this move changes nothing about existing engagements by Microsoft projects, why is the company doing it? I see little evidence that the hostility to open source has softened at the executive level, though Ballmer no longer derides open source openly. But on the ground, the market is forcing Microsoft's hand.

Given Microsoft won't discuss the rationale for forming its new open source subsidiary, I can only speculate on specific motives. Here are four possible explanations.

  1. Consolidation. The new subsidiary takes the standards and open source teams that already exist at Microsoft and groups them under a better name than "interoperability." Interoperability may sound good to some audiences, but it reeks of "our way or the highway" and "isolated competing teams" to lots of developers and has little resonance in the world of open source where words like "contribute" and "participate" have more meaning. This new subsidiary has Paoli's existing Interoperability Team at its heart, now nicely rebranded. I also hear that the former Platform Strategy Group -- responsible for engagement with Linux -- has been folded in to Microsoft's cloud operations.
  2. Create a career path. Jean Paoli was one of the people named on the original XML specification and has been a loyal long-term player in the standards arena. High-powered standards wonks are like spymasters: necessary to their political bosses but with work that is usually better unseen. The result is a lack of career path in many corporations. This move could be a reward for Paoli and his team, recognizing his long-standing loyalty to the company. It's a charming and plausible tale if true, although it seems like an expensive and complex way to retain staff.
  3. Firewall open source licensing. Licenses like the GNU GPLv3 are an inescapable fact of open source, and they do a fine job protecting their communities. Indeed, if Google had used the GPL version of OpenJDK in Android, it might not be in court today facing Oracle. However, the licenses do that by placing responsibilities on corporate participants, especially on how they handle patents. Most modern licenses include a "patent peace," removing rights from community participants who turn out to be patent litigators. They also secure explicit or implied blanket patent grants from contributors in most cases. A separate subsidiary provides Microsoft an "arm's length" relationship so that license terms can't affect them directly, especially so any implied or explicit grants to rights from their extensive patent portfolio are limited.
  4. Firewall patent liability. After years of scepticism, Microsoft became an ardent admirer of software patents -- particularly their earning potential -- when embedded in both de facto and de jure standards (a technique the company gained when it hired the "inventor" away from IBM). Today, Microsoft is actively lobbying for RAND (arbitrary fee-based patent licensing) terms to be allowed in standards so that those "standing on the shoulders of giants" have to pay the giant. But there's a corresponding risk of taint from others planting these "patent bombs" in standards. By having the work done in a separate company, Microsoft limits the liability it faces from those playing the same game by seeking fees and alleging plagiarism.
1 2 Page 1
Page 1 of 2