I'd like to offer a few thoughts of my own on the relationship between the mandatory curve and exam-writing -- with the necessary caveat that I have only graded one set of exams in my brief academic career.
At Georgia, professors are required in 1L classes to turn in grades with a mean GPA between 2.9 and 3.1. In 2L classes, professors have to turn in grades with a mean between 2.9 and 3.2.
In my evidence class, which had 115 students, I reserved the right in my syllabus to increase grades for exceptional participation. What that meant was that I graded in two steps. First, I had to blind-grade all of the exams, curve them, and submit them to the registrar with anonymous numbers instead of names. The registrar noted the blind grades and returned them to me with student names next to the grades. In the second grading step, I then adjusted individual grades to reflect class participation and returned them to the registrar.
For both grading steps, I had to stay between the 2.9 and 3.2 mean. That meant, in order to adjust grades upward for class participation, I had to make sure my initial grades were not too close to 3.2 (which was my target going in to grading; I'm a softie at heart). As it turned out, my initial grades meaned at 3.05, which allowed me to adjust somewhere around a dozen grades upward one level (B to B+, for example) without exceeding 3.2.
In terms of how the mandatory curve affected by exam-writing, I think it encouraged me to write a very difficult evidence final -- one hour of multiple-choice, two hours of essay. I knew the exam was very difficult, but I also knew that it would produce a much broader range of raw scores than an easier test. If the test had been easier, the scores would have been much closer together and breaks between grades would have been very close to arbitrary. As it was, there were natural distinctions between grade levels, and I was able to use participation, when deserved, to make sure that one point didn't mean the difference (pardon the pun) between a B and a B- or a C+ and a C.
In all, I'm quite satisfied with how my first encounter with grading turned out. I had a few students come to me to discuss their grades or complain about how I worded the essay questions, but the majority seemed satisfied with their grades.
Any professor that claims grading is an objective process, however, is lying. Even with a seven-page answer key for the essay questions that listed 145 possible points, I still had many difficult judgment calls to make. Was sentence X close enough to point X on the answer key to deserve credit? Did I give points for information that was not explicitly mentioned by the student but that she had to know in order to provide the answer she did? How many points should I give for a particularly creative answer? For a particularly well-structured one? Trust me, these questions plagued me over the five days that I graded exams.
The old adage is true: law professors are paid to grade exams. The rest of the job they do for free.