> However, it's recently been suggested that management might begin to use this timesheet data for other purposes, such as assessing individual programmer productivity and performance and for payment of bonuses.

That strikes me as a rather ineffective way to measure 'performance'!

Say you write a fantastic piece of code in 10 minutes, when you were expecting it to take an hour. Using the performance=hours worked metric, there's an active disincentive to move on to the next challenge for the next 50 minutes! You end up with a company paying programmers to read slashdot/PM/whatever (not that any of us would ever consider doing such a thing ;-) ).

It does raise the question of how one might measure the productivity of coders. Sometimes the most productive thing to do is delete everything and start from scratch, sometimes things take less or more time than expected, you know the score...

So is there a way to measure productivity that can be generalized for most programmers in most environments, or is it something that we just can't measure meaningfully?