Debugging time should be measured

UPDATE: Changed title from “The dangers of debugging“ to “Debugging time should be measured“ because it's a mistake to give the impression that cleansing a computer program of ones bugs and defects is an evil act.

One of the things that I'll be trying to do next year is to encourage other developers that I'm associated with to spend less time "on the debugger".  I read a couple of blog entries this morning which extrapolated out some of the facts and figures associated with debugging; one of particular interest was this one: Cost of Debugging.

My preference for a development day is something more like this:

• first 3 hours PLAN (Work out how to do it)
• next 3 hours IMPLEMENT THE PLAN (Code it)
• final 2 hours FIX THE IMPLEMENTATION (Debug/Test it)