I recently ran into a curious set of circumstances and wanted to compare notes to other people. Over the last 6 years my managers have opted several times to take the "expedient" way around a problem rather than truly stopping to solve it. This seems to be an ingrained habit amongst managers at my organization. Yet we have enough CM data that I can go back and data mine to see how many of these decisions led to subsequent problems and how much time we spent on the subsequent problem. So I did.
First the facts. In 17 cases where serious design issues were noted after an implementation had been done and for which refactoring was highly recommended, in every single case more time was spent chasing bugs related to the expedient decision than was initially estimated as it would have taken to fix it.
Further, I can note that the developers on our staff have a pretty good track record of meeting their estimates for implementation tasks - to within 10% of the intial estimate over 90% of the time - so I have good reason to trust the developers on these items.
In the lowest boundary case, total refactoring time was estimated at 4 man-weeks and the time spent chasing bugs was 5 man-weeks (so far). In the upper boundary case, the total time to refactor an existing problem was estimated at 3 total man-months and time wasted due to this problem has exceeded 33 man-months.
Given this information, I asked one of our managers today why he was not going to fix problem X the recommended way. His response was that neither I nor any of the rest of the staff could quantify how bad the problem really was. My reply was that no, I cannot tell you how an anti-pattern is going to bite you in the ass but I can tell you that every single time you've done this in the past, you've lost the time game. Every single time. In other words, you've gone out in the rain 17 times and gotten wet, yet you are going to sit here and tell me that maybe, just maybe if I go out this time I won't get wet?
But the manager would not budge. He literally refused to consider data about his organization, his people, and his process right in front of his face. He wanted the scope of the this specific problem quantified, which I cannot do.
Has anyone else experienced this? And if so, how have you handled it?
Previous Topic
