In my presentation (PDF, ODF, PPT) to an audience of open source ISVs contemplating relationships with Microsoft – the long time standards bearer for proprietary software – I likened the relationship between the firm and the movement to the Odd Couple. Though I declined to cast the respective parties.
After the past few days at NXT and OSBC, if anything, I’m more aligned with that viewpoint than I was when I put the deck together.
As some of the attendees of the Microsoft Technology Summit have noted, Microsoft is making some genuine, good faith efforts to evolve its attitude about and work more effectively with open source. From Apache to Eclipse to Mozilla, Microsoft is working – and working effectively, by most accounts – with some of the more important open source projects on the planet. Projects, notably, that in every case compete directly with Microsoft products.
Personally, I consider this evidence that Sam Ramji and his team, like Jason Matusow and his team before him, are fighting the good fight with some success. While opinions remain divided within Microsoft, those working towards better relationships with open source appear to be winning over fearful executives who’ve managed to convince themselves that open source was everything from communisitic to cancerous. Which can’t be, I think we’d all agree, an easy job.
Who would have imagined, I asked the audience at NXT, a future in which Microsoft would publish – on a Microsoft web property no less – articles such as “Installing Apache on Windows” and “Recovering Data from Windows systems by using Linux.” What in the Get the Facts is going on? Two or three years ago either of those articles would likely have been every bit the firable offense that discussing M&A activity out of order would be.
Inevitably, opinions vary, but I don’t think it’s entirely naive to hope that Andi is correct when he says “I believe Microsoft has finally understood that their closed nature has significantly hindered the growth of their eco-system.” It’s optimistic, perhaps, but not naive. Or at least not entirely so.
When Ray Ozzie says, “by increasing the openness of our products, we will provide developers additional opportunity to innovate and deliver value for customers,” I’m sure that on one level he’s simply being genuine, and recognizing a business reality obvious to many of us. But on another, I’d bet that he’s eagerly anticipating the very same future that Tim O’Reilly warned of. A future in which software is increasingly, albeit it not exclusively, consumed via network services. A world in which the license applied to that software – and the closed or open nature of that software – is less relevant than it is today.
Until that day, which is frankly not within hailing distance, Microsoft must contend with a world that places a much higher premium on heterogeneity than it once did. Than it did in Microsoft’s pre-Google apex. Which means working more effectively with other software, and open source software in particular.
It took a while, and was not the smoothest transition imaginable, but from all appearances, Microsoft does want to work with open source. Sure, they’d probably prefer to kill it, but even the most optimistic ‘Softie has to admit that that’s simply not in the cards at this point. And if you can’t beat them, well, you know how the saying goes.
Aside from the aforementioned work with Apache et al and the NXT conference, Microsoft sent its general counsel – who presumably has other things he could be doing with his time – to speak at OSBC. And they’ve invited some smart folks – smart folks who disagree with Microsoft, or so the claim goes – up to Redmond for the purpose of listening.
But we all know which road is paved with good intentions. The question that remains unanswered for me is this: can Microsoft ever truly work with open source – downstream and all – without significant compromise on its principles around intellectual property? The principles that Smith called yesterday, the “courage of its convictions?”
Based on what I’ve seen, I suspect the answer is no.
They can partner effectively with open source ISVs; JBoss, MySQL, Novell, SugarCRM and Zend are sufficient to prove that to my satisfaction. With Apache, Mozilla and others, Microsoft is also demonstrating a newfound and welcome ability to collaborate strategically with projects. All of which is good. At least, if you care about customers.
However well they can partner with such entities, however, there are three obstacles that will have to be addressed before Microsoft can work effectively with open source in all of its myriad forms.
The Cathedral Preference
In the Q&A yesterday afternoon, Smith allowed that while open source innovation and leaders may come from many quarters, the crew from Redmond was generally more comfortable engaging with the cathedral than the bazaar. He was referring, of course, to Raymond’s seminal essay, but my immediate reaction was instead geopolitical. Microsoft, I believe, is going to have to learn the same lesson that post-Cold War Western governments are currently struggling with. That the days of negotiating with leaders are nearing an end; that the future is negotiating with populations.
Microsoft is far from alone in their need to digest this message; few if any vendors do this well at the moment. But those with longer and deeper ties to open source have been quick to grasp, as an example, the idea that all members of a population are not created equal. Committers, like elected representatives, carry special weight and unique privileges, which is why we’ve seen and will continue to see acquisitions that based more on commit rights than actual technology.
It will be interesting to see if Microsoft, with its obsessive commercial focus, can adapt to a more distributed world. Can Microsoft, in other words, learn from Kilcullen?
The Commercial Distinction
Last month, Microsoft chose to honor my birthday by releasing a wealth of APIs, documentation, protocols and specifications under liberal licenses for the first time. Or maybe the date was random. Anyway, as Smith noted, the immediate reaction to the release was not generally the acclaim that presumably was hoped for, but immediate and near total confusion.
To be fair, this is not really Microsoft’s fault. Open source licenses such as the GPLv3 are at least as complex as the terms of Microsoft’s patent and open specification promises. As I’ve discussed previously, given the often permanence of legal promises, contracts and licenses, it’s not unreasonable for an organization the size of Microsoft to proceed carefully and craft the language carefully. Far from it.
But what quickly became apparent when Sam attempted to explain the implications of the patent promise to the audience at NXT was that Microsoft’s reluctance to let go of the commercial distinction is a problem. Maybe an insolvable one, in the absence of substantial compromise.
Within a few minutes, the audience began to ask questions to which there were no readily available answers. This was due not to any failure on Sam’s part, but the simple fact that the commercial distinction that Microsoft insists on drawing is intrinsically incompatible with the philosophies of open source licensing generally, and likely the Apache 2.0, GPLv2 and v3 licenses specifically. When code is released under the Apache license in particular, it must be free of precisely the sort of entanglements and restrictions the patent promise imposes.
I am certainly no lawyer, so it’s possible there are non-obvious loopholes or workarounds, but I’m not hopeful.
While Microsoft has indeed been more open than at any point in its previous history, then, it is still not open enough to be compatible with a variety of open source licenses, and therefore projects. Which is unfortunate, but unavoidable from where I sit.
The Patent Affection
The commercial distinction mentioned above, of course, can be traced entirely back to Microsoft’s views on its patent portfolio. To its credit, Microsoft has backed patent reform legislation, but the Redmond giant will simultaneously state unequivocally its belief that software patents are both necessary and beneficial for innovation. Beliefs that I, for the sake of full disclosure, do not share.
Irrespective of my personal opinions on software patents, however, there is a legitimate question to be asked of Microsoft: what is behind the defense of software patents? Is it emotion – the same emotion that’s slowed an acceptance of open source – or strictly business?
Microsoft is quick to point out that the royalties they demand for patent licensing are low; several percentage points below the EU recommendations, in fact. But this discount only deepens the mystery of Microsoft’s current strategy vis a vis its patent portfolio.
No one questions the need to acquire software patents, to be sure. Whether you love the current system or hate it as I do, it will punish vendors that do not acquire software patents at least for defensive purposes. Even vendors that profess to not believe in software patents confess to a need to accrue them.
But what is the particular benefit of maintaining programs to license them and monitor for infringement? If you’re IBM, and your patent licensing revenue stream is eight figure dollar amounts, the explanation is obvious. But Microsoft’s patent revenue stream is dwarfed by IBM’s, from what everything that I’m told.
In which case, why incur both the financial hit of the licensing program and the public relations hit that comes with licensing or threatening litigation around patents?
Microsoft has a standing annual legal bill, we’re told, of ~$100 million dollars to defend itself from patent related litigation. Unless the licensing revenue stream easily eclipses that amount, why is the current system worth defending? Why does Microsoft insist on speaking out in defense of a mechanism that appears, if anything, to negatively impact its shareholders? It seems like a flawed equation.
Asked precisely this question yesterday on three separate occasions, Microsoft’s Smith returned polite but less than substantial answers. When I asked specifically about the economics, his response was that it implicitly proposed a false free/expensive dichotomy that Microsoft did not and could not subscribe to.
But frankly, that seems like so much sophistry to me. While it’s true that cheaper patent licenses are obviously preferrable to expensive patent licenses, that assertion fails to address entirely the argument that a less restrictive approach to its patent portfolio would offer substantial innovation and PR benefits.
To be honest, I can’t help but conclude that Microsoft’s longtime adherence to the doctrine of intellectual property value is coloring its behavior with respect to software patents, to the firm’s possible detriment.
In the Meantime
While it’s true that I believe Microsoft is unnecessarily handicapping its efforts within the open source space, however, I do believe it’s important to note the progress made. I could trot out the cliches about journeys of a thousand miles or the time to construct Rome, but I suspect you get the point as well as I do.
Microsoft is not today as open as I would hope, and that may or may not be the case for you. But I’m hopeful that the trajectory they’re on now will yield positive returns in future, particularly if some of the risks they’re taking now produce positive returns.
There are many whose opinion of Microsoft was irrevocably stained by ill considered comments made years ago. Which is each individual’s right. For my part, however, I echo Sam’s comment:
When I joined IBM, it was the undisputed evil empire. Here’s hoping that Microsoft makes the journey successfully.
If any of the discussion above helps them navigate there more quickly, I’ll be happy to have done my part.
Disclosure: Microsoft is a RedMonk client, as are Eclipse and IBM. Apache and Mozilla are not clients.
