I attended the kickoff of the
Linux Foundation's Collaboration Summit yesterday, at Google headquarters in Mountain View, CA. Also representing the OSA was Kumar Thiagarajan, VP of Community Dev for Adaptive Planning.
Overall it was a productive and enjoyable event, and it achieved the Linux Foundation's goal of bringing together a critical mass of Linux thought leaders to collaborate and agree upon Linux's future directions. There was also lots of customer representation as well - I met with Kenneth Topp from Goldman Sachs, Michael Breuer from NYSE and the always irrepressible Tim Golden from Bank of America, among others, and judging from
the attendee list, many more were in attendance. Hats off to Jim Zemlin and team for pulling together this event, and at relatively short notice, from what I understand.
Indeed, this Summit had a "revival" feel to it, with Jim giving a motivational sort of keynote, highlighting the success of Linux thus far, but also pointing out the work that remains, and imploring the community to work together. "It's easy to point out philosophical differences and poke holes at each other's ideas, but as leaders of the community, we must work on bringing people together and be the seekers of truth, not finders of contradiction." This message was then reinforced by an excellent (albeit shortened) talk by Mark Shuttleworth about how we can take better advantage of collaborative tools and techniques to work together better, not just within individual community project "silos" but also sharing knowledge across communities as well. Additionally, a series of panel discussions expounded upon how the collaborative spirit could address other challenges in various areas (kernel development, customer adoption, application ecosystem growth, and legal/licensing issues).
My own takeaway is that these issues aren't just limited to the Linux community, but the broader community of open solutions as well. Actually there were two moments in particular that struck me as us having a lot in common, once while talking to customers, and once while listening to the "application ecosystem growth" panel session.
The Customer Perspective
The first was during lunch, when I was fortunate enough to share a table with B of A's Tim Golden and Michael Breuer from the NYSE. I asked the open-ended question of what, if anything, was getting in the way of driving further adoption and getting further value from their technology investments, and inevitably interoperability came up.
Michael had an interesting take: Early adopter customers like the NYSE have been deploying and integrating both packaged and custom code for a while, so they understand the issues. The problem is when he tries to feed back to the vendors. First, every vendor's interpretation of "interoperability" presumes they are the center of their respective universes, with little movement towards truly common approaches. Proprietary and open vendors are equally guilty of this, but perhaps for different reasons: Many proprietary vendors are proactively executing competitive differentiation and lock-in strategies, while open vendors end up preferring to just sticking with what they know - their own "silos" of expertise. And second, there are just too many vendors that need to be influenced. "It gets tiring having the same conversation over and over again."
He was referring specifically to system-level interoperability, such as networking and device drivers, and consistent approaches to systems management and monitoring. But the issues are more general than that, and would apply to code at any level in the stack, up to the applications layer. Getting things to work together, and being able to "partner" with the vendor community to make this less onerous, is critical to customer success.
Tim Golden summed up this problem well during his panel session later in the afternoon: "Is the 'community' spelled with a big 'C' or a small 'c'? It feels like a loose federation of interests, while big customers want predictability over time. I can't produce predictable results over time. If I have an app that depends on interoperability of OS, device drivers, UI, peripherals, etc., and there's no natural inclination for those constituents to work together to help me deliver my app, then I can't promise predictability to my business users."
The Application ISV Perspective
Another panel featured rock stars of the open applications tools and infrastructure market, including Darren Davis of Novell, Mike Milinkovich of Eclipse, Kay Tate from IBM, Scott Nelson from Real Networks, Brian Aker of MySQL. They were asked about the application ISV perspective, namely what gets in the way of ISV's developing applications on Linux.
The responses varied, from Mike's assertion that ISV's will develop to whatever platform has broad adoption (with the implication that Linux has more work to do here, at least on the desktop), to the costs of certifying applications on Linux. This latter point led to spirited discussion, because it pointed to the lack of coordination between relevant projects. Scott made a good point: "It's not just testing on the kernel. Do I have an installer? What patch/update mechanisms do I need to support? How does logging and management work?" With so much to certify against, the lack of a single certification program is a barrier to many ISV's. LSB is a good start, but needs to expand.
Indeed, Scott's point is true, but he could have gone much farther. Kumar later told me that he needed to be restrained from going up to the mic and yelling "Guys, wake up! It's not just about the OS!" Later that evening, we had good conversations regarding this with Jim, and also Amanda McPherson, the Linux Foundation's director of marketing.
There's Much More We Can Do
There's definitely more we can do together to provide a unified front to application ISV's, and also to end customers. We don't need to do this in a centralized way. That's the proprietary answer. Instead we can do this through collaborative efforts like what the Linux Foundation is doing at the systems tier with LSB, and what the OSA is doing at the applications tier.
We all have to, and we will, keep making progress on this. Customers are feeling the pain today. ISV's are feeling the pain today. And our proprietary competition is taking notice. Squabbling within the community doesn't help, it just feeds their FUD regarding the open community's ability to deliver. We all believe that open and collaborative behavior will consistently deliver better results than closed and unilateral behavior, but we need to keep working at it. There is more, much more, that we can accomplish together. "Be seekers of truth, not finders of contradiction" is a message that couldn't have come at a better time.
