I think the right answer is, "As much like a bazaar as it can be," and therein lies the weasel-like nature of my answer. A standard-- and to some degree, valid-- complaint against the stereotypical libertarian is that the stereotypical libertarian would be perfectly okay to bring back the days of the travelling snake-oil salesmen. Because, hey, customers should educate themselves, they can always sue for the cost the medication if it's nonsense, and surely consumer groups can be created to provide education, right?

Stereotypical libertarian, I said, which is about as common in reality as the ultra-far-left liberals who really are communists-- the last time I saw either one in the wild, it wasn't really in the wild, it was in college. No serious person I know wants that, nor the abolition of the FDA for either food or drugs. Reform, perhaps (and fairly modest reform at that) but not abolition. A complete lawless jungle of a bazaar is not the answer.

Even in the context of pure information technology, my instinct is to say, "as open as it can be," and weasel out again. Because on the one hand, I firmly believe that open source has a large place at the table, in the future. On the other hand, it's a lot easier to open source a generic application like a word processor or even an operating system-- everyone who programs understands those applications. It's a little more difficult to open source serious industrial and research applications just because there are a lot fewer people who know both the software engineering and the application space... and those people tend to want to be paid. Highly. We don't open source air traffic control software, after all. (Well, we talk about it but I remain skeptical.)

Now, the Google article is very interesting, because they talk about both open source, and open standards. Google is obviously in favor of open source, because they don't make their money selling product, they make their money selling advertising. Backing open source makes perfect sense for them. I'm not sure how well the model works for medicine-- most people are pretty paranoid about their medical information, and Rosenberg's model of Interest Based Advertising is going to run into some pretty tough ethical and legal challenges as they apply to health data.

On the other hand, open standards are extremely important. Right now, there is barely even a notion that multiple medical devices attached to a single patient should even know about each other, much less communicate with other, so data communication standards for medical devices are practically non-existent. This is barbaric, but at least it means we can start from the ground up and design them instead of having three device cartels fighting each other.

Likewise, the idea that information should be able to flow efficiently from hospital to hospital seems to be a lot more popular outside hospitals than inside hospitals.

Here, the government can probably do a great deal, with comparatively small investments. There is no fundamental reason that a stethoscope should not contain an RFID or barcode reader to read the badge of a patient, should not have an 8 GB memory stick inside to record the audio and other relevant data, and should not have a bluetooth or other wireless connection to dump all the information to a central storage location at the end of the day. Yeah, $500 today, but $50 in five years.

More importantly, there is every reason to expect that that stethoscope, from any vendor, should work here in Illinois, or there in California, or even way down there in Florida. There is every reason to expect that a hospital in Texas should be able to request those files from a hospital in Minnesota, get them in a matter of minutes, and have their devices make sense of them.

Every piece of medical equipment should be expected to do this. (Not to mention, when a stethoscope like that is fifty bucks, I'll buy my own. When they're five bucks, everyone will wear them as pendants and stream the information to their desktops at all time.)

It really goes beyond that, though-- not only do we need open standards in devices, we need open standards in research models, so that advances from one research group can be transferred to other groups-- but more importantly, can be moved farther down the product pipeline faster.

I see I have stunned the community into submission, with my long commentary.

What you are suggesting appears to me to be inevitable, but it is only a part of what we will see in the next decade in terms of health care delivery as information technology and biotechnology merge.

I also believe that for present the "World Market" can only be a "World Bazaar", that Intellectual Property and Patents will very quickly become dead issues, because they simply will be impossible to police.