To me the funniest part of the novel Snow Crash, by Neal Stephenson, was his description of the internal management practices of the Feds.
In the fictional world of that novel, most centralized states had collapsed, and the territory of the former United States was home to dozens of competing networked “government” franchises. The Feds, or the former federal government of the United States, was one of those competing governments (although it claimed continued jurisdiction over the former territory of the United States). Its main source of revenue was software design for private clients.
From the way the Feds organize their software design operations, they seem to have read The Cathedral and the Bazaar, recoiled in horror, and decided instead that Brazil was the way to go.
Everybody’s assigned their tiny little share of the project on a need-to-know basis, with their individual pictures of the project resembling that subcommittee of a subcommitee Winston Smith sat on to decide whether the Eleventh Edition of the Newspeak Dictionary should put braces inside brackets or vice versa. Smith’s assignment was actually a model of transparency, in comparison, because at least he knew it had something to do with the Newspeak Dictionary. The overall design of the Feds’ software, or even its basic purpose, is outside the scope of anyone’s need-to-know below the highest level. And nobody can alter a single line of code without reference to endless policy manuals in three-ring binders; what’s more, since these policy manuals are revised every few weeks with endless interdeparmental meetings, most of the new code written has to be thrown out every time the policy is changed.
The coder’s first order of business, after clearing the hurdle of urine tests and personality profiling to get to work, is to spend until noon or so reading all the interdepartmental memos on new regulations or changes to the existing rules for writing code. Most of the afternoon is spent rewriting the portions of code rendered obsolete by changes in the rules (with none of the hundreds of programmers working on any project having any idea what it’s actually for, of course—that’s classified).
Even the interdepartmental memos include suggested reading times, with the surveillance system monitoring compliance. Anyone who scrolls through in less than the suggested time lacks proper respect for the importance of policy memos, while anyone who takes too long is suspected either of incompetence or of taking an unauthorized bathroom break. And anyone who reads it in exactly the suggested time to the second is a smartass who needs attitude counseling.
I’m not sure who the customers for the Feds’ software are supposed to be, but I get the feeling the IT department at my employer (and probably yours) would be among them.
Until last week, I thought Stephenson’s farce—hysterically funny as it was—was a grossly exaggerated depiction of even the worst real-world bureaucracies.
But no more. According to an op-ed by Jonathan Vaccaro at the New York Times, it takes 96 hours after the Taliban arrive in an Afghan village for an Army commander to secure the necessary approvals to act. The company in which Vaccaro was embedded failed to interdict the Taliban in some 70 percent of cases because its commander failed to get the required eleven approvals in time. Travel in anything but a 20-ton mine resistant vehicle requires “written justification, a risk assessment and approval from a colonel, a lieutenant colonel and sometimes a major” (over half the villages in Afghanistan are inaccessible to such vehicles). The Taliban walk in or ride donkeys.
The bureaucracy runs to the highest echelons. Small aid projects require endless delays for approval (the opening of a small free health clinic was delayed eight months after it was built “while paperwork for erecting its protective fence waited in the approval queue”). While Taliban propaganda operations turn on a dime in response to events, “our messages have to inch through a press release approval pipeline, emerging 24 to 48 hours after the event…” Battlefield commanders are required to submit reports in PowerPoint, “with proper fonts, line widths and colors so that the filing system is not derailed.” So, um, if you could put the new cover sheets on the T.P.S. reports, that would be great, m’kay?
John Robb, who blogs at Global Guerrillas, makes a couple of points about the American military’s organizational model.
First, “risk mitigation trumps initiative every time.”
Second, rather than using new communications technology to “enable decentralized operation due to better informed people on the ground,” the military instead uses it to “enable more complicated and hierarchical approval processes—more sign offs/approvals, more required processes, and higher level oversight.”
Just another example of why state capitalism is doomed. Small, agile, bottom-up organizations will eat government and corporate bureaucracies alive. One of my favorite sayings is that the twentieth century was the era of the large organization; by the end of the twenty-first, there won’t be enough of them left to bury.