Yostivanich

Caper Jones, in an unpub­lished 1977 study for IBM, found that the very best devel­op­ers are much more pro­duc­tive than the worst pro­gram­mer — when work­ing on small projects. The best devel­oper will com­plete a 1k line of code (LOC) effort 6 times faster than the lousi­est. The pro­duc­tiv­ity delta falls to 2x on a 64k LOC project. Beyond a few hun­dred thou­sand LOC both sorts of peo­ple per­form equally well. Or equally poorly.

via EETimes – The leg­end of the super­pro­gram­mer. I think our ten­dency is to stick awe­some pro­gram­mers on a large task and hope they can lift it up by them­selves instead this argues for stick­ing them on small projects, on the large project they may be more valu­able on a very small sub­set but on the whole just as impor­tant or un-important as a nor­mal programmer.