From John Robb, who seems to have coined the term "open source warfare":
[The Iraqi] insurgency isn't a fragile hierarchical organization but rather a resilient network made up of small, autonomous groups. This means that the insurgency is virtually immune to attrition and decapitation. It will combine and recombine to form a viable network despite high rates of attrition. Body counts - and the military should already know this - aren't a good predictor of success.
Given this landscape, let's look at alternative strategies. First, out-innovating the insurgency will most likely prove unsuccessful. The insurgency uses an open-source community approach (similar to the decentralized development process now prevalent in the software industry) to warfare that is extremely quick and innovative. New technologies and tactics move rapidly from one end of the insurgency to the other, aided by Iraq's relatively advanced communications and transportation grid - demonstrated by the rapid increases in the sophistication of the insurgents' homemade bombs. This implies that the insurgency's innovation cycles are faster than the American military's slower bureaucratic processes (for example: its inability to deliver sufficient body and vehicle armor to our troops in Iraq).
Second, there are few visible fault lines in the insurgency that can be exploited. Like software developers in the open-source community, the insurgents have subordinated their individual goals to the common goal of the movement. This has been borne out by the relatively low levels of infighting we have seen between insurgent groups. As a result, the military is not going to find a way to chop off parts of the insurgency through political means - particularly if former Ba'athists are systematically excluded from participation in the new Iraqi state by the new Constitution.
Third, the United States can try to diminish the insurgency by letting it win. The disparate groups in an open-source effort are held together by a common goal. Once the goal is reached, the community often falls apart. In Iraq, the original goal for the insurgency was the withdrawal of the occupying forces. If foreign troops pull out quickly, the insurgency may fall apart. This is the same solution that was presented to Congress last month by our generals in Iraq, George Casey and John Abizaid.
Unfortunately, this solution arrived too late. There are signs that the insurgency's goal is shifting from a withdrawal of the United States military to the collapse of the Iraqi government. So, even if American troops withdraw now, violence will probably continue to escalate.
What's left? It's possible, as Microsoft has found, that there is no good monopolistic solution to a mature open-source effort. In that case, the United States might be better off adopting I.B.M.'s embrace of open source. This solution would require renouncing the state's monopoly on violence by using Shiite and Kurdish militias as a counterinsurgency. This is similar to the strategy used to halt the insurgencies in El Salvador in the 1980's and Colombia in the 1990's. In those cases, these militias used local knowledge, unconstrained tactics and high levels of motivation to defeat insurgents (this is in contrast to the ineffectiveness of Iraq's paycheck military). This option will probably work in Iraq too.
In fact, it appears the American military is embracing it. In recent campaigns in Sunni areas, hastily uniformed peshmerga and Badr militia supplemented American troops; and in Basra, Shiite militias are the de facto military power.
If an open-source counterinsurgency is the only strategic option left, it is a depressing one. The militias will probably create a situation of controlled chaos that will allow the administration to claim victory and exit the country. They will, however, exact a horrible toll on Iraq and may persist for decades. This is a far cry from spreading democracy in the Middle East. Advocates of refashioning the American military for top-down nation-building, the current flavor of the month, should recognize it as a fatal test of the concept.
For me, this is as interesting for its flat assertions about the nature of the Open Source ("[F/] OSS"] movement as it is for his clarification of the term as it applies to warfare. There's some very interesting -- perhaps revealing -- language, here. I can remember reading John Robb a few years back, but I don't remember anything in particular that made him stand out from the other tech-bloggers I was reading at the time. Here, he's saying some things that are different, that not everyone else (in the tech-blogging "community", at least) is saying.
For example, he's acknowledging the success of IBM, and how they got it: By 'letting the enemy win,' or more precisely, by buying the enemy their uniforms. IBM spends a ton of money on Open Source development. No other company with the arguable exception of Google has as strong a claim in Open Source councils.
Another example: While he seems to praise with one hand, he does something very interesting by tossing IBM into the same metaphorical stew with the right-wing Salvadoran and Colombian militias, trained to do the nastiest kinds of dirty work by our own CIA at our own School of the Americas. Folks at IBM who get the allusion might well be pissed off by it; I expect it's intended not as an insult, but rather as a precise analogy. The analogy bears expansion, though, because most Americans are woefully ignorant of their own history -- especially the small and dirty parts of it like what the Salvadoran militias (and, hell, their regular military) actually did to their own people, with our help and encouragement. If Robb is right, we're in the process of doing something very similar, again, and this time on a far larger scale.