Local Maxima

My friend Swathi blogged about her frustrations at having to handle trivial bugs filed by testers with the sole intent to increase the bug count.

While I understand her furstrations towards such testers, I think the real problem might be her manager.

I assume that the eventual goal of her manager is to get the best quality work from his colleagues. To achieve that he chose to reward people for what he thinks would improve quality:
1) He penalises developers if the bug count is high
2) He rewards testers if the bug count is high
Hence, he creates an artificial tension between the developers and testers, in the hope that it will improve quality. I have seen this being employed in atleast couple of projects; and I think the tactic is doomed.

When management decides to reward something which they think will improve quality (in this case, bug count), people will soon figure out a way to maximize that thing – without getting anywhere near the intended result. This is called local maxima and I have read references to scientific studies which prove this phenomena (too lazy to dig out references

Joel Spolsky mentions the futility of encouraging increased bug counts in his series on management techniques. Joel must have much knowledge from the field on this – his company produces a popular bug tracking tool.

The Expert Mind

Interesting article from Scientific American. The article summarizes studies made by psychologists on chess masters and tries to explain what expertise really means. The article is long, but was worth reading for me.