Thursday, July 28, 2005

Wednesday, July 27, 2005

The desire to seek an explanation is part of human nature. We accept “just so” stories to avoid uncertainty but we pay a price because the explanations come in lieu of understanding and often lead us astray.

Today's (July 27, 2005) Boston Globe has a story on George Gilder and his advocacy of Intelligent Design. The argument is classic – if you can't explain something you have to ascribe it to something or someone. The Intelligent Designer is the explanation of last resort. Malcolm Gladwell (Blink) gave a talk at last year's PopTech in which he noted that if you ask people for an explanation they will give you one. In his example people were asked why they preferred one chair over another. But the stories had no predictive value – they were made up because it's in our nature to fill in the holes. I think of it as the cognitive equivalent of the persistence of vision. People who are missing part of their visual fields are often not even aware of it because they cover it so well. In fact, we all have a physical blind spot in our eyes but it's hard to see what we can't see!

The bell curve is a very interesting phenomenon that seems to defy explanation. Look at how carefully it is laid out – each ball falls into just the right place. There's also “action at a distance” – each ball that falls on one side has an exact counterpart on the other. “Obviously” such a system cannot arise naturally because it requires that every aspect be perfectly coordinated.

Stop now – what is wrong with this argument? Can you explain the bell curve any better? Can it “just happen”? Can it evolve out of more basic steps?

Of course not only can it but it does. IBM had a nice exhibit at the 1964 World's Fair (which has been distributed to museums around the US including the Boston Museum of Science) in which balls are dropped onto a series of pegs and form a bell curve.

The bell curve emerges before your eyes out of the independent behavior of each ball. If you have one peg and drop balls half would fall to one side and half to the other. If you have three pegs in a triangle half would go to one side or other for each peg. At the two lower pegs the balls fall to the outside or inside. They have different outsides but share the inside – those more balls would pile up on the inside. The result is a bell curve but the steps are obvious. As you add many balls it smooths out into the nice bell curve. There's nothing to prevent all the balls from falling to one side – it's not impossible but unlikely enough to be treated as impossible.

If you look at each ball or peg or component the bell curve doesn't exist. It just emerges from the dynamic behavior. “bell curve” is just a name we give to an observed phenomenon.

The danger in positing a designer is that it denies us the ability to explore and learn more. It seeks to explain away mystery and leaves us at the mercy of an inscrutable designer whose motivations are inscrutable. It leaves us afraid of the unknown rather than confident in our ability to live with uncertainty. It leaves us intolerant of those who threaten to unravel the web of stories that shield us from the wonder of the world.

Tuesday, July 26, 2005

What does Digital Rights Management have to do with evolution? DRM is a way of assuring that the “content owner” can maintain control. That seems innocuous in itself but it has the effect of limiting the marketplaces' ability to change. This makes sense in limited cases as it allows investors to recoup the cost of their investment and make a profit but if DRM works too well it prevents growth. A marketplace is a dynamic system that keeps changing. Why doesn't the marketplace simply devolve into chaos? The reason is that it is an evolutionary process – one that provides opportunity for creating new results. We can think of this opportunity in terms of Chris Anderson's long tail – it represents the value to be discovered rather than what is obvious.

Marketplaces that work can capture the results that are viable while surviving those that don't work. They renew themselves dynamically. Without this process of renewal marketplaces stagnate and fail. While the goal of DRM may be noble, if taken too far it leaves us impoverished.

I was annoyed and angry to find that I couldn't use my high resolution monitor to watch HDTV content. Instead I am supposed to buy an HDMI compliant monitor that would be more expensive and less capable than the one I have. For some reason even after my attempt to use an “unauthorized device” the program guide information is only available on the component (three wire) and not the composite (single wire, standard TV) output. Not only am I not allowed to choose how I want to watch, I am at the mercy of a set top box that is befuddled by the complexity of implementing the scheme!

Something is very wrong. While Microsoft may consider itself only helping out by providing facilities to aid and abet such stifling control they are doing damage by thwarting the dynamics of the marketplace. Sadly, both Microsoft and Intel seem to be determined to undermine Moore's law by saddling it with fatal complexity in the hope of insuring their incumbency and the incumbency of other industries that are past their prime.

Tellywood is defined by the asymmetric control afforded by older technologies. It is intent on keeping this control even if it means we cannot do anything for ourselves in case they might not capture all of the value of their works and in case others may create competing works.

The desire to reap the rewards of ones efforts is understandable but we must have a balance. Such control must not come at the price of denying others any control at all and must not come at the price of preventing economic growth.

Imagine where we would be today if Edison were able to maintain sole control of the “moving picture” technology. He maintained stifling control until his competitors decamped to Hollywood where they could assert their own stifling control.

Microsoft is going to prevent what they call “hardware attacks” (as well as “software attacks”) on premium content. Such attacks include what others call fair use. My attempt to watch content on my own screen is an example of just such as “hardware attack”.

From http://www.microsoft.com/whdc/device/stream/output_protect.mspx

New output content protection mechanisms planned for the next version of Microsoft® Windows® codenamed "Longhorn" protect against hardware attacks while playing premium content and complement the protection against software attacks provided by the Protected Environment in Windows Longhorn. These output protection mechanisms include:

• Protected Video Path - Output Protection Management (PVP-OPM) makes sure that the PC's video outputs have the required protection or that they are turned off if such protection is not available.

• Protected Video Path - User-Accessible Bus (PVP-UAB) provides encryption of premium content as it passes over the PCI Express (PCIe) bus to the graphics adapter. This is required when the content owner's policy regards the PCIe bus as a user-accessible bus.

• Protected User Mode Audio (PUMA) is the new User Mode Audio (UMA) engine in the Longhorn Protected Environment that provides a safer environment for audio playback, as well as checking that the enabled outputs are consistent with what the content allows.

• Protected Audio Path (PAP) is a future initiative under investigation for how to provide encryption of audio over user accessible buses.

Microsoft and Tellywood are working to assure that you can't buy a monitor better than a dinky Tellywood-approved monitor that matches their narrow vision of the future. An entry from engadget http://www.engadget.com/entry/1234000727051424/ points out that the next step beyond DVD protection is the use of revocable keys. I posted an essay RIAA Plans to Sue Hearing Aid Manufacturers to make this point but it was satire – unfortunately it may be far too close to the truth.

For me the issue is not so much whether people can choose to protect the content but the effects on innovation. This is a Tellywood that would have totally and utterly defeated VHS and you wouldn't be able to make home movies. And they wouldn't have gotten huge new markets. The inability to choose your own LCD screen creates a huge barrier between computer screens and TV screens. The whole silly idea of TV screens being at six feet and computers at two feet is one of those silly Power Point inspired theories that is at odds with reality.

Microsoft and Intel seem to think it is in their interest to cooperate with this approach and limit the ability of users to find new opportunities. The long tail gets truncated. It's like Cisco helping China control the spread of “bad ideas”.

It's useful to read some recent discussions on Dave Farber's mailing list for context (http://www.interesting-people.org/archives/interesting-people/200507/threads.html). In particular the discussion on science education and the Gates' contribution to the Discovery Institute which advocates Intelligent Design. This doesn't mean that Bill Gates necessarily endorses their theories but the question of intelligent design is an important issue in how we view marketplaces.

Too bad evolution is taught in biology class because it makes it hard to see the larger issues. Dynamic systems are evolutionary systems and if you try to limit their dynamics they fail. If you believe in intelligent design you can assume that systems can be guided. Marketplaces are just complex systems. If you give the incumbents the role of the intelligent designer the systems will fail because you don't allow for new ideas.

It's easy to convince oneself that things are indeed working well and we shouldn't risk tampering with it. A good (or, perhaps, bad) example is the Bluetooth protocol. It demos well but if you try to use it in anyway that is not anticipated it will fail as David Berlind points out in http://blogs.zdnet.com/BTL/index.php?p=1634. The Dial-Up Networking (DUN) Profile builds on the idea of a circa 1980 Hayes Modem and then adds more layers in an attempt to pile Internet protocols on top of it. With enough effort you can make it work – at least for a while.

I was struck at “D” by Intel's Paul Otellini's approach to the complexity of networking. His solution is to add another layer of mechanism. That only compounds the problem and makes the system more perverse. The whole point of the approach I took to home networking was to reduce the amount of basic mechanism rather than piling more on it. David Berlind observed one consequence of the “pile on” approach in trying to deal with Bluetooth's “DUN” (Dial Up Networking) protocol which adds connectivity by going all the way back to the Hayes modems and then adding networking as a special case on top of all the other mechanism with all the baggage.

I've also started to listen to IBM's Irving Wladawsky-Berger describing IBM's autonomic (self-healing) computer and their grid architecture. Here too complexity is “solved” by adding more mechanisms.

For those who believe that one can predict the future and that the world is organized into nice hierarchies it makes perfect sense to add mechanisms to the pile and leave it to the prescient of incumbent business to define and limit us to that future. It confuses business with marketplaces.

For those who recognize a rich evolving world such efforts to limit opportunity do far more damage than just deny us the ability to innovate. It makes it very difficult to use what we have because the only combinations that work are those that are anticipated. We get the kind of task-oriented design that gave us Bluetooth. It's what Microsoft uses in trying to “improve' the user interface in their systems.

It's also the womb-to-tomb misinterpretation of end-to-end that Bill Gates expresses at “D”. Microsoft is trying to do us favors by providing us with a complete solution rather than one that is open to allowing us to take different approaches.

This why I keep emphasizing that teaching evolution in biology classes leaves us without understanding that evolution is a characteristic of all systems not just “special” ones. Without such understanding it is difficult to see how and why the Internet works. Even more to the point why it works despite and not because of governance. Why complexity is an emergent property of the lack of understanding. We don't “solve” complexity by layering on top of it. When we design systems we have to go underneath the system expose the simplicity.

It's not at all fair to accuse those who thwart marketplace processes as being “anti-evolutionists”. Even though I think it is obvious the onus is still on me to demonstrate that the mechanisms are the same. I still claim that there is a basic philosophical alignment akin to the one that George Lakoff posits in Moral Politics. It is hard to trust the marketplace because at any point in time it's too easy to see the “right” answer. It's even more difficult to see the importance of these dynamic processes when cling to the present for safety.

It's like looking at the weather. You can't just see that it's 28º (Celsius, Fahrenheit, take your pick). You have to look underneath at the dynamic behavior. The same is true for marketplaces – what you see is a result of a dynamic process. If you try to legislate against change you don't even get to keep what you think you have.

Marketplaces don't just work but are necessary. We can frustrate them for short periods – the US Constitution grants only limited exceptions. But only at a price that increases rapidly over time.

I better post this piece soon because I'm required to acquire the rights for each word lest the coiner assert ownership …