PenguinPolitik: Only Ballmer could go to Linux
Well, okay, I can appreciate the position of conventional Open Source ideologues on Microsoft as much as anyone else. I have a lot of them as friends; I drink and break bread with them on practically a weekly basis. But as someone who lives in both of these worlds and feels comfortable in either of these circles, and as someone who could have gone either way in their career, my take on these things is somewhat different. I have been a close observer of Microsoft for many years, and this image of this evil, calculating machine looking to dominate the world is somewhat overblown. It's all too convenient to put Microsoft or even a guy like Steve Ballmer into a simple box. I've met the man and spoken to him on a number of occasions -- admittedly it's been a good ten years -- and I have a hard time believing he is the completely unreasonable and monstrous human being that the Open Source faithful want to make him out to be. Bombastic and loud? Frequently talks before he thinks? Sure. I can name half a dozen prominent technology execs that have similar qualities. Monster? Nah.
Click on the link below for more.
Glasnost is the term that I used to describe the practice of what I believe Microsoft, or at least certain factions of the company are currently engaging in. There are guys like Sam Ramji and members of his team who genuinely want to make Microsoft products interoperate with Open Source and want to put as many pet projects on the Open Source Lab's table as they can - and as far as the group is concerned, there are no "boundaries" that define what those projects should be. When I posited the question "well, I can see you guys assisting SAMBA because the EU has enforced it, but would you consider independently helping the WINE project so that Microsoft Office and other key WIN32 applications ran near flawlessly on Linux and UNIX?" the answer was "Yes, we'd consider it." So I think there is clearly some room for negotiation and productive discussion.
There is another word that comes to mind for this type of productive discussion, and that word is realpolitik. It is defined as politics or diplomacy based primarily on practical considerations, rather than ideological notions, and it tends to be goal oriented. We cannot expect to have any reasonable dialogue with Microsoft as an Open Source Community unless there are mutually achievable goals in mind of a specific nature -- only then can we move into more complex, ideological negotiations. Realpolitik was a practice that was attributed by most historians to Otto Von Bismarck, minister of Prussia (also coincidentally, born on April Fools day) who unified Germany during the 19th century. However, perhaps the most well-known example of Realpolitik was the opening up of China to the West by U.S. President Richard Nixon. Hence the famous phase, "Only Nixon could go to China."
This gets back to the much maligned Steve Ballmer. Sure, Ballmer is a hard-lined anti-Open Sourcist, who has compared the development model to Communism and flat-out intellectual property theft. Like Nixon, who was as hard lined on the Communists as they come, I can't think of a better guy to be an ambassador to the Open Source Community and someone who could begin multi-party talks in realpolitik-style goal-oriented negotiations. He is, at the end of the day, a businessman, and will do what makes sense for Microsoft from a business perspective.
So let's talk about some goals, stuff that we know should be simple to achieve and quite easy for Microsoft to justify and goals that the Open Source community could quickly realize if there was mutual cooperation.
First, would be to create a development kit or a standardized Open Source development platform for Windows so that Linux and UNIX applications could run natively and with good performance. Microsoft has a platform, and it's called SFU (Services for Unix) formerly branded as INTERIX. The technology, an enhanced POSIX layer for Windows, was acquired as part of the Softway Systems purchase in 1999. I asked Sam Ramji if this is something Microsoft would consider releasing under an OSI-approved, GPL-compatible license, and as it turns out, this is one of things that they are seriously considering as a positive gesture to the community. The implications of such a measure are huge, because it would greatly improve the performance and ease of development of POSIX and cross-compiled Linux and other FOSS applications on Windows.
The problem is that SFU lacks a complete GNU toolchain. Well as it turns out, someone has a complete GNU toolchain for Windows, and it happens to be Microsoft's arch-enemy, RedHat, who's Cygnus IP and Cygwin project, would make an excellent marriage for such an Open Sourced venture. Hello, Mr. President, are you ready for your China trip? If Microsoft would see INTERIX released as some sort of foundation -- in a model where they are a participant, mentor and sponsor -- but not a controller, such as Sun's position with OpenSolaris and OpenOffice -- we could very well see Windows as becoming one of the best Open Source development platforms there is. Popular platform equals more Windows licenses and an increased Microsoft developer ecosystem and more services sold. Developers, Developers, Developers, Developers. I'm pretty sure Steve Ballmer understands that.
Another nice project for Microsoft to be engaged in would be to release the complete specifications to the NTFS and DFS network filesystems under an OSI-approved license. Sure, now we got SAMBA, and all the protocols that go with it, and an eventual pure compatibility goal with CIFS and the latest SMB2 spec, but Windows's native filesystem? Nope. The NTFS-3G kernel drivers that community developers have reverse engineered are functional, but they are hardly high-performance and I wouldn't trust their ability to maintain data integrity on an NTFS volume with a ten foot pole. A fully released NTFS specification would take the voodoo out of Linux/UNIX and Windows file system interoperability and would provide for a native, open and industry standard high performance journaled file system that would open up the utility, security and data recovery and defragmentation utilities and services market to Linux, areas which are currently sorely lacking. And with a fully implemented NTFS, it will be much easier to sell Linux to end-users on the desktop, because it's a file system they already know how to deal with.
I'm sure I could think of a whole bunch more goal-oriented PenguinPolitik, but I'm running out of space. What easily definable goals do you think the community and Microsoft should set? Talk back and let me know.