Perhaps the issue here is our over-centralization and reliance on tech built by big companies. As has been demonstrated by both IED makers that figure out how to confound countermeasures in 5 days and by our troops who created improvised armor for Hummers in the field (or is demonstrated by Linux development), the best way to develop cheaply and quickly is to pool knowledge. If there aren't websites on how to foil new IED designs as they come out, there should be. Our troops should be communicating with each other to find ways to disable and destroy these devices. A company's approach is to take an existing problem and build a solution, but if the problem moves the company can't keep up (like windows and it's endless patches against viruses). Decentralized networks work better in this instance.

"the best way to develop cheaply and quickly is to pool knowledge"

I'll pick on this one because its the easiest to take down. Now, mind you a lot of this is a subjective argument. Nothing is done in a vacuum, that needs to be recognized by me, but at the same time I can't think of a single earth shaking development that was done through collaboration.

Steam engine, Archimedes screw, TNT, Theory of Relativity, Theory of Gravity, UNIX (arguable since it was invented by 2 people and the product testing was done by university students), airplane, automobile, assembly line implementation, Bessemer process, the list goes on and on. Now mind you I realize that other people are involved in a products implementation - but the key ideas were the product of a single entity or small organization. Even Linux was largely the product of one person - as evidenced by the continuing existence of FreeBSD, a product that has been around LONG before Linus Torvalds ever became known.

As for Linux being an example of cheap and quick deployment of solutions. As someone who does architecture design for my company I can say that Linux was never an example of that prior to it being bought up by the big corporations. All the interesting developments of Linux have occurred since taken over by companies. Novell, SGI, IBM, HP, RedHat have taken the core product and turned it into something useful for the corporations - whereas before it was only a lab platform. The OpenSource community could never develop efficient MPI libraries to match those produced by SGI and IBM. Without SGI and HP Linux would never have run effectively on the IA-64 platform. It took the integration of their proprietary platforms to make Linux something a corporation can use. In fact it is the OpenSource community that is holding back the full development of that platform - I see it every single day at work. I don't recommend Linux when a proprietary is available because of the poor communication and coordination in the OpenSource community. Not when I can get a platform that is extrodinarily well documented, tested, and supported 24/7 by a Fortune 500 company. When we do run Linux, we never run untainted kernels - otherwise memory allocation, hardware support, MPI code, and a host of other optimizers would never see the light of day.

Now, don't get me wrong. Lots and lots of excellent ideas have gone by the wayside due to the originators lack of human skills. Collaboration is a good thing in that it moderates the inventor and steers him/her/it toward a better concept. But ultimately the idea came from one person, it took a committee to hatch the idea to fruition.

Damn, we agree again!

In the meantime, why don't we blow up every cell tower near where an IED has gone off. If there are people who need cellphones perhaps they will inform on IED's before they go off. If not, at least cellphones won't trigger another IED in that neighborhood for a very long time.