Using software metrics for managing developers is usually frowned upon. That's too bad because they sure make software management fast and easy - as demonstrated in the video "Metrics-Based Software Management - A Hands On Approach".
Note: This post, and the associated video, are meant to ridicule the idea of "Metrics-Based Software Management", NOT endorse it. A few too many people are taking it seriously. For my position on software metrics please read Testivus on Metrics.
My previous Artima post, "Software Metrics Don't Kill Project, Moronic Managers Kill Project", caused quite a stir. The split in the comments made it clear that there are still some people who doubt the legitimacy and usefulness of software metrics.
I can understand lazy and incompetent developers being afraid of having their work measured objectively. But, for the life of me, I can't imagine why someone with the superior brains and skills required to rise from a lowly programmer to the position of manager would not want to use metrics. Who are these metricophobics? Methinks that they are just spineless namby pambies who can't handle the truth and are afraid of having to take decisive management action based on actual facts and figures.
If you are one of those jelly-fish managers who indulges in touchy-feely employee reviews and gentle feedback based on abstract and highly-subjective criteria, you need to shape up or get out of software management altogether. We don't need you, and we don't want you!
How do you shape up? You can start by watching, and learning from, a real software manager in action (i.e. me).
In the following video, you can see how I use the time-tested C.R.A.P. software metric to handle a developer who refuses to play by the rules.
Watch it and learn. CAPISCE!
If you are ready to become a real manager, or a responsible developer, download Crap4j from Crap4j.org, start to use it, and stop making excuses.
>>>> But if a metric relies on statistics that have been widely criticized in the literature as invalid and misleading, the name (CRAP) is well deserved. >>>>
The two metrics we use in the first iteration of CRAP, cyclomatic complexity and basis-path coverage, have been the subject of many studies and peer-reviewed papers some of which may be critical, but many of which are supportive. These metrics are neither ideal nor perfect but I, and many others (both in industry and academia), have found them to be VERY useful (both as a developer and as a manager) when properly and judiciously applied.
Of course, the key words are "properly and judiciously applied". I am aware that metrics can, and are, more often than not, misused, abused, and over-used. As a matter of fact, this satirical post (and video) were meant to bring attention to the problem of metric abuse.
You make a lot of interesting and valid point in your first post. I apologize for not replying sooner, but you invested serious time and thought on that post and I needed to find enbough time for a worthy response. It's obvious that you care about the topic at least as much as I do. It's also clear that we don't agree on a few points; but I see a lot of common ground and I look forward to discussing it more and benefit from your knowledge and experience (as I have learned from you in the past).