Where have all the good geeks gone?

For those who may not know, i work for a big company. My livlelyhood depends almost entirely on the fact that the overwhelming majority of people in IT at a big company are somewhere between incompetent and outright dangerous. In fact, entire careers, and by extension lives, are built upon this enormous, precarious scaffolding.

That said, as an organization, we continue spending tons of resources recruiting and hiring "quality" employees. i often wonder how so many people can work so hard at something for so long, and still have the result be almost a complete failure.

Is it the nature of software jobs that organizations are unable to locate qualified personnel?

Most professions have clear indicators of competence that can be quantified. For a carpenter, a room should be built square, the roof should not leak, and so on. For a printer, the copy should be aligned properly on the page, the ink coverage should be uniform and match a known sample.

Not so for the IT professional.

Fulfilling the documented business requiremets has almost no relationship to the quality of the software, design, maintainability or cost. These systemic quality aspects are virtually impossible to quantify in any meaningful way. Every system built is so different from the one before it that comparing any two systems is a lot like trying to compare the mass of a blue whale to the price of gasoline in the midwest.

What's the point? Are we doomed to hire lousy programmers and carry them around as dead weight until they retire?

i suggest a new approach. An approach employing the concepts of lean manufacturing and the well proven fact that great programmers are orders of magnitude more productive than mediocre programmers.

Keep on hiring, but start firing. Aggressively. Let go of the dead weight and backfill aggressively. Instead of viewing a project as requiring a ten person team, recognize that the project could be done better and faster if you had three talented developers. Now, still fund and staff for ten, but aggressively remove all but the top two on a regular basis. Cycle the bottom 80% out every four to six weeks. Eventually, after several cycles, you will find that you simply can't get it down to only 2 people. At this point, you have the optimal team to complete your project. Save the top, drop the rest, and drive through to completion.

Radical, yes. Heartless, perhaps. Pragmatic, maybe not. Effecive, definitely.

Think about it.