EC ruling: careful what you wish for

Unless you have been living under a technology rock, you know by now that Microsoft lost its antitrust appeal. The European Court of First Instance rejected most of Microsoft's arguments. The court also ruled that they EC went too far by appointing a Microsoft-paid "special master" to oversee Microsoft's compliance, something some pundits tried to spin as a "win" for the company.  I fail to see why. That legal footnote sounds like a technicality, a bit like saying prisoners don't have to pay for their own prison guards.

My thinking on antitrust has evolved considerably since I first broached the subject seven years ago. My response to the American ruling is more reflective of my current thinking on the subject, though today I would say that the conduct prescriptions crafted by Kollar-Kotelly aren't just "acceptable," but are actually quite useful.

The EC also had some conduct regulations, chief among them rules that force Microsoft to document core protocols used in its desktop and server operating systems. In truth, I think Microsoft should have created this documentation irrespective of whether or not antitrust regulators felt it necessary to force them to do so.

I don't think it makes ANY sense for large companies to play lock-up games with protocols anymore. The days when companies or individuals would feed their critical data into black boxes is long past. Computing is simply too critical an aspect of our daily lives, and it is that valid principle which gives rise to demands for more open systems in general, and the open source / free software movement in particular. Every movement is a reaction to something, even if some of its practitioners take their principles to extremes.

Being open doesn't really hurt Microsoft, in my opinion. At a simple level, it is a great internal motivational principle. No company should want its employees to feel "safe" from competition, and historically, Microsoft has done its best work when faced with long competitive odds.

Documentation requirements, however, are likely to spread Microsoft technology to places where Microsoft on its own could not spread it. This is a good thing for Microsoft, as it creates interesting new upsell opportunities in future. That's the "careful what you wish for" aspect of this blog. The EC has just ensured that Microsoft technology will have a much broader reach than Microsoft could ever manage on its own. Will UNIX fans find that they are encroached from ever more sides by Microsoft technology now that the EC has forced Microsoft to make those protocols widely available for everyone to use?

Those are the good aspects of the ruling, though my attraction to them has little to do with whether or not they will actualy change Microsoft's market share. I think they make sense technically-speaking, reducing unnecessary competitive barriers as they benefit the software art. They may or may not change the balance of power in the industry, however, particularly if that dominance has something to do with the peculiarities of software markets.

This is where my "death to antitrust" side comes out, because regulators faced with the fact that their policy prescriptions aren't creating market conditions as predicted in their antitrust labs often do less than rational things.

For instance, take restrictions that mandate that Microsoft must offer a version of Windows free of Media Player. Clearly, this is a restriction that has almost zero effect, as Windows version N has practically ZERO sales. Granted, Windows version N costs the same as the version that ships with Media Player, but are regulators now to decide the "relative value" of new features in an operating system (if you are nodding your head in agreement, then you are now so far off the path of economic rectitude that the concept of "path" no longer exists)?

Besides, the principle underlying this prescription is a bit goofy. Regulators are basically saying that computer evolution should stop at a certain point, mandating that end users be responsible for the addition of new features whether they want to or not, or else allow OEMs to randomize things such that end users aren't sure what base level features they will find on their new computer. One advantage to integrated features is consistency...just ask Apple. Apple's entire "computer as a TV set" business model, where systems loaded with Apple-only features is considered a competitive differentiator, would be torpedoed if they weren't allowed to control integrated features.

Likewise, end user applications are different than the API guts that support them, and I don't think regulators want to rip that out unless they truly want to freeze operating system evolution in amber. When Microsoft first created Windows version N, they ripped out not just Media Center, but the whole media pipeline. That caused Real Networks to scream, as they relied on that pipeline even if they didn't use the user interface wrapper that Microsoft offered in the form of Windows Media Player.

Operating systems will need the ability to offer certain base level services if they are to continue to serve as platforms that evolve to support future needs. That means that messaging architectures, user presentation renderers (e.g. HTML rendering components, even if you don't have the frame in the form of Internet Explorer), media pipelines and other things will need to be included in an operating system if programmers who write atop them are going to be able to have a consistent baseline upon which to build their applications.

That, at least, is the way I see things. The EC, however, doesn't appear to see things that way, and may be using a simple results-based analysis to determine if their policy prescriptions are working. That's a problem, as it's a simple fact that for most of computing history one company has tended to dominate. That makes sense if you understand how software works. Creating a consistent base for application development is hard, but the market naturally wants that as having one creates economics of scale and lowers costs dramatically. Hence, one company tends to dominate, and if it wasn't Microsoft, it would be someone else.

I accept that that dominance can have some negative consequences. Requiring protocol documentation restrictions helps to reduce those consequences, as does preventing big companies from using their market power like a weapon (e.g. restrictive contracts that lock out competitors).

Regulators, however, have to know their limits. They have to understand that their job isn't to shape markets like artists sculpting clay. They can deny large companies the ability to avoid competition through restrictive contracts. They also can ensure that competitors have the information they need to compete.

Beyond that, however, is dangerous economic territory. If regulators knew when to stop, I would be a more enthusiastic supporter of antitrust. Unfortunately, history has shown that regulators rarely know their limits, and the EC has said little that convinces me otherwise.

But I hope to be proved wrong.