The classes I have taken in graduate school have all been "seminars". These generally take the structure that [appears to] gives the professor no duties beyond selecting papers and assigning grades. When I first came to grad school, with the ink still wet on my college diploma, I was thrilled. The "senior seminar" I took the last semester of college was of the same format, I had loved it (although perhaps that was due to the home-baked treats the professor's wife provided each week), and I was looking forward to more of the interactive, discussion-based classes rather than being lectured at.
Now, I am enrolled in one of the peculiar mixed undergraduate-graduate classes. Graduate students are held to higher standards and have more assignments than the undergraduates, but the class time is the same. I find myself absolutely thrilled with the experience. We sit in rows facing the projector, rather than in a half-circle facing each other! We get out of the classroom after just an hour twice a week, instead of after three hours straight once a week! I get to be taught by an actual professor, rather than by my equally-inexpert fellow graduate students! I am asked specific, factual questions rather than being asked to pontificate on about my opinion! Most importantly - most refreshing - we are discussing basic concepts rather than the pluses or minuses of broad theories or specific experimental design!
Not all the changes are positive, of course. The student's work in a seminar class is usually to read the papers, participate in class and "lead" the class discussion once or twice a semester. For the degree-requirement classes there were usually some form of reading reaction and a final paper, but the "upper-level" seminars don't usually bother. This mixed, undergraduate-style class had me spending a good five minutes entering the various reading reactions, homework sets, and larger assignments into my calendar. Barely a week will pass without me having to turn in a reading reaction or homework set. My field is not one generally given to "homework sets", except for the basic stats class, so that's a novelty in itself.
I can certainly understand why professors prefer to minimize course preparation if at all possible. But if the real reason this course seems so wonderful is just the change of pace from the classes I've been taking the past few years - as the seminar did at the beginning of graduate school - then I have to seriously argue in favor of offering graduate students more classes instead of seminars. Even the professor seems far more engaged in this environment than in the traditional "seminar" I took with him last year. The enthusiasm for a different style of learning is going to carry over into whatever subject matter is presented, with the potential of making me fond of a subject matter when I only enrolled in the course because it's related to my advisor's other line of research and meets some degree requirements.