Alex writes about our shared experiences at IBM.com in the 1995-1997 era while commenting on an essay by Clay Shirky. I'd add that one downside to the ad-hoc, small group developed software is that it didn't always scale on a variety of axes (sometimes traffic, sometimes amount of data, sometimes maintenance). Too often the executive overlords would see how cheaply some solution had been developed and then extrapolate unreasonably that no futher investments of time, people, or other resources would be necessary. When the I/T BT/CTIO octupi struck, they struck at the weak parts of the system (security, reliability, scalability) even though frequently the "formal" organizations were no better at these aspects than the informal groups.