Slashdot Mirror

So, pretty much this.

And, yes, I agree. Metrics and "objective measurement" are a plague on business, effectively treating all employees as inherently untrustworthy while incentivizing grossly degenerate gaming of the system. Knowing, recognizing, and cultivating trusted, intuitively competent people cannot be replaced by the commoditization of labor via the adoption of metrics. Conversely, reliance completely on "gut feel" is far less effective than inventive intuition combined with empirical evidence.

Solution? Cultivate people who know how to do the job right and know when to override your metrics. Don't rely on the system alone to deduce who is doing well when. Make sure that managers talk openly and regularly with employees 2-3 layers down the hierarchy.

All of this is more effort, so screw it. Let's just make a new process and require every employee to go to training about it.

Any discussion of Agile and its value calls for a mention of Orson Scott Card's How Software Companies Die. Card makes the very valid point that good software developers abhor deep and thorough control of the "process" by which they make software. Typically, the developers I've known to like Agile are the mediocre ones.

Agile won't make a bad team good, and it will make a good team a bit slower, with sprint synchronization, a meeting-heavy culture, and commoditization of development resources. The only big win that I see for Agile is the ability to more quickly call the failure shot. There's value to this, but every single Agile group that I've encountered, worked in, and heard of has eventually moved to valuing the process over the practice of developing software.

Believe that being able to measure the process is critical to producing good software? You're a manager who has read too many crappy management books and spent too little time in software development. Believe that being able to measure the process is more important than what comes out of it? You should take a look at Agile.

You should try to read How Software Companies Die by Orson Scott Card. A short essay on the same subject as GP. Really nice.

self-organizing, that is...

http://www.zoion.com/~erlkonig/writings/programmer-beekeeping.html

Cheers!

  -Captain Fairly-Obvious.

Or how about just linking to the article (here) instead of making us jump through hoops?

I like this article by Orson Scott Card titled "Why software companies die". It's really short (and really old - 1995) - go read it.

I like Orson Scott Card's take on this issue he's talking about programmers rather than sys-admin's but I think there are similarities.