Metrics — Sean Kelly: "I myself follow the Personal Software Process (PSP), a system developed at Carnegie-Mellon, to gather such data and track it over time. As a result I can usually estimate the time/cost/defects of my future projects with good accuracy; after all, we are creatures of habit, and it takes quite a bit of change to effect change in the quality of our products. (As an aside, my productivity from C++ to Java tripled; from Java to Python tripled again!) Yet if you lack such data, then making these guesses will be quite challenging.
Still, there's no time like today to start gathering metrics. Try building some smaller software systems, then look at the time it took you, and the size of the programs in lines of code (lines of code, or LOC, is a terrible metric for program complexity, but no one's yet come up with a better one, and as Gibbs once said, 'Everything can be measured in some way that is better than not measuring it at all.'). For your ultimate project, look at aspects of smaller projects that are similar to what you want to accomplish."
Keine Kommentare:
Kommentar veröffentlichen