Agreed. "Experienced" seemed a dubious definition. Experience != Skill, as well.
One of the things in the paper was that there was a population with middling errors that did "many" builds, and a bifurcation of 2 populations that did "few" builds - one with few errors, and one with many errors. (pg 9, section 4.1 "How often do builds fail?")
Could it be that there is a group of gurus who don't need to constantly build, and which batch up testing of larger feature sets, and another group of perpetual freshmen who spend so much time fixing their fault riddled crap that they simply cannot submit very many builds? Yes, I know they attributed at least some of this to the employee role being done, but it makes me wonder. "Experience" would not tease this pattern out. (while it seems to me customary to build more often in the initial stages of using a new language, we've all worked with "that guy" who just bumbles through, year after year)
One of the things in the paper was that there was a population with middling errors that did "many" builds, and a bifurcation of 2 populations that did "few" builds - one with few errors, and one with many errors. (pg 9, section 4.1 "How often do builds fail?")
Could it be that there is a group of gurus who don't need to constantly build, and which batch up testing of larger feature sets, and another group of perpetual freshmen who spend so much time fixing their fault riddled crap that they simply cannot submit very many builds? Yes, I know they attributed at least some of this to the employee role being done, but it makes me wonder. "Experience" would not tease this pattern out. (while it seems to me customary to build more often in the initial stages of using a new language, we've all worked with "that guy" who just bumbles through, year after year)