I was thumbing through Joel Spolsky’s new book on hiring talented programmers at the bookstore the other day. He asserts in the first chapters that great programmers are ten times more productive than bad ones. Could that be true? I played around with git today to generate statistics to see how it looks on our refactoring project this year:
user commits files changed LOC added LOC deleted
dev0: 453 4258 +81538 -100344
dev1: 152 2297 +54462 -45343
dev2: 104 550 +24779 -17475
dev3: 82 250 +6465 -7852
dev4: 73 209 +3920 -2782
dev5: 56 208 +5158 -10049
dev6: 24 116 +3297 -7823
Wow. Of course KLOC isn’t everything, but take it on faith that dev4 causes more problems than he solves (dev5,6 were only on the project a short time), and that dev0 is the most awesome programmer that writes on this blog. And I thought surfing the web 7 hours a day was too much…
Thanks, although it is more of an indictment of those who went before. WTFs galore.
Now if only the new code will work as well as the old...
Ick, reminds me of the days of having to use PVCS. Actually everywhere I worked that used such horrid commercial tools (i.e. ones for which the CTO hopefully got kickbacks), developer revolt ensued such that the "official" repo was only where someone occasionally uploaded snapshots from the real repo. Sort of how I'm now treating our svn repository now...
Ginormousdatabasecompany probably has something neat, though, with lasers and such.