I feel that grades should be an objective measurement of a student's performance that the student can use to measure his or her own progress rather basing the grade off of how one compares to other students.
There are some classes in which you can do this and some you can't.
I had a politcal theory class that was very difficult. The teacher presented four or five philosophers and you had to make them speak to each other in an essay test. You had no idea what the questions would be.
Something like, "How would Rousseau respond to Bush's war on terror?" "What would Machiavelli do?" What would Foucault say to Machiavelli?
Well, I'm not sure how you grade that on a point scale. While your answer certainly could be wrong or right, it had to make sense no doubt, but it's bit hard to deduct points or give credit for something like that.
What our professor did was read all the papers and get an idea where they were from best to worst and lay the grades out that way. I had no problem with that. While one could make the case that you could be giving worse grades to great papers, I would say the chances of that happening are slim. If you take a group of 50 students, they are usually going to peform the same way. You'll have two or three that will knock it out of the park, you'll have some above average, and then you'll have the average ones with some stragglers on the down side of things. There is no way that 40 people will be writing A material. If they are, the class probably isn't challenging enough.
I've always been told that if a professor turned in grades to the department and everybody got A's, his class was too easy and would have to be made harder.
My experience with college, though, was that most classes were as tough as they needed to be. Not many had curves.
Only the ones that were out of sight hard had good curves. That allows the teacher to not compromise the material just because most people don't understand it.
How do you dumb down Astronomy? You either get it or you don't.