He would run it on the one dataset he likes to work with. If it worked at all, then he'd come wanting to show me (and I'd always praise him...) and bugging me about what to do next. I'm too nice and was too scared of making him unhappy and having him quit... So I never pressured him to do the miserable task of running his code on all the other datasets... And often his code didn't work great on the first dataset anyway. For example, we might want to identify all the "features" in an image but his code identifies maybe 30%. But he'd be so stubborn about the algorithms and packages that he likes to use that I got worn out and stopped pressuring him.
No vesting. I own what I own, whether I stick it out or quit...
|