Tuesday, March 5, 2013

Do high benefit/cost work!

Tom DeMarco wrote an article a while ago: “Software Engineering: An idea whose time has come and gone?” (2009)  I guess I am just catching up on my reading.
One topic is: are metrics useful.
I think the real issue is: how useful is it that we ‘control’ projects?
His comparison is:
1. If we have a project that probably will cost $1 million and returns a benefit of $1.1 million, then we must control cost carefully. Or the project will be underwater.
2. If we have a project that probably will cost $1 million and returns a benefit of $3 million, then some variance in cost is not important.
His contention (and mine) is that there is plenty of work in category 2.
This does not mean have no budget and do no planning. It just means that we should not look at planning and control as the main drivers of success. The main drivers are: identifying the problem well, and delivering a truly innovative solution.

No comments: