News for hackers. Stuff that matters.
Re:Size is the best predictor of code quality by Confusion on Monday September 26, 2011 @05:06AM (#3038144)
Another reason (in addition to those already mentioned) why these numbers are not easy to interpret:

  small routines [..] had 23% more errors per LOC
If the smaller routines had, on average, 100/1.23 = 81% or fewer of the lines of the larger routines, then they still had fewer errors per routine.

There are no comments on this comment yet.