[cmalloc]

I don't think the grandparent's point is an issue of sides. It's more of a problem of gathering a useful heuristic of an engineer's performance. Given a variety of every day engineering tasks, how can you know exactly the difficultly of a person's work short of having done it yourself? I believe that experience is the answer in some cases, but others it is not. As another poster stated, it could also be done by working with the engineer through it all.

[humanrebar]

That's basically my take. A lot of evaluation is properly subjective and qualitative. Both bosses staring at deadlines and union negotiators clamoring for paper trails and predefined metrics are going to be bad for the industry and therefore workers, at least in the long run.