Busting the Myths of Programmer Productivity

Are the great programmers really 10 times faster than the rest? What does this difference in productivity even mean? What productivity distribution should we expect between professionals? How can we use this knowledge? In this webcast, we make the most of a large set of programmer training data using repeated measures to explore these questions.

What attendees will learn:

• For routine tasks, professional programmers have a narrower range of productivity than we first supposed, but almost half of the variation in individual productivity is noise, making programmer rankings suspect.
• Rather than finding the “fastest” programmers, we should find competent people and give them the training and environment they need to succeed.

Speaker: Bill Nichols

#software #coding
@Software Engineering Institute | Carnegie Mellon University

[SKARTHIKSELVAN]

Valuable talk not only for sw engineers but also for hiring managers.

[peterfireflylund]

I have looked at the assignments his stats are based on. They are *very* simple, short, and limited in scope. Examples: make a program for counting the number of logical lines of code in a program (that is, omitting comments and blank lines) or a program that performs a chi-squared test of some numbers (using a previous assignment about creating a numerical integration program that uses Simpson's method). The requirements are very detailed and low-level (so the hard part is already done) and the assignments are meant to take less than a day to write (a couple of hours at most, I think).

In other words: not very realistic, to put it mildly. And *nothing* that warrants the conclusion of a "busted" myth -- if anything, I'd say this is evidence that the actual spread is likely to be *big* under realistic conditions when it is already so big (2x-5x) under these unrealistic conditions.