Forgive if someone has already stated the obvious, but the difficulty of rating the teacher profession, as opposed to other professions, lies in the intended outcome. For an engineer, there are many objectives metrics to measure code work against. The same is true for many non-programming corporate jobs (usual metric is $$$).
But what is the intended outcome of teaching?
Better test scores? (Google Scholar 'standardized testing and success' and take your side of the argument, but at best the jury is still out)
The three R's? Social development? Curiosity for learning? All of the above?
I understand your underlying point, but it is dangerous to contrasts teacher evaluations with other professional evaluations, lest we find ourselves programming students a science) and not teaching them (an art).
Crazy idea alert! With intended outcomes varying so much for teachers and teaching environments, I'd offer that teachers should be only evaluated by fellow teachers, parents, and students themselves.