Steven Conn's Rise of the Helicopter Teacher from The Chronicle of Higher Education:
A week before the first paper was due, a young woman in my class raised her hand and asked where the rubric was.
Shamefaced and stuttering, I had to admit that I had no idea what a rubric was. She helpfully explained that this was a set of guidelines explaining what I expected them to write, how I expected them to write it, and how each aspect of the paper would be evaluated. A set of boxes that students could check off to guarantee that they had met my expectations. For all intents and purposes, in other words, an outline for the paper.
Except a rubric isn't an outline for the paper. Not really. A rubric--or maybe I should say a good rubric--is a set of expectations for an assignment.
When we give an assignment we have a mental model of what a good paper looks like. We also know what excellent and crappy papers look like. A rubric simply takes our mental models and puts them on paper so they can be shared with students.
Yes, you're telling the students what an "A" paper looks like. But your exceptations already existed before you made the rubric. It seems unfair to penalize a student because they failed to read your mind.
Are rubrics perfect tools? Of course not. Can they be overwritten and limiting? I guess so.
However, we've probablty all had the experience of saying "I have no idea what this teacher/professor wants" when working on an assignment. I simply can't understand how that experience enhances student learning.
(p.s. My favorite variation of bad rubric is just a bunch of generic nonsense like "paper is proper length" that wastes everyone's time.)