Due to the sudden influx of "not that bad" ratings, let's not forgot that's "he's not good at all," either.

His project organization is terrible. His lectures are mediocre. Rohr seems like a nice guy though.

Based on the prior reviews, I came into Rohr's class completely terrified. Having received an A in his class, I will say that there is only partial validity to the reviews.

Teaching:
It is true that Rohr is a lousy teacher. He is a very boring lecturer. He covers a lot of material that is nothing more than reference in the book, and does not do a good job of differentiating what is important and what is an unimportant detail. I find the historical parts interested, since Rohr has seen it all and frames thing in the context of the evolution of computers. The good news is that the class is 80% programming and arithmetic, so lecture is only minimally important and being able to write code is MIPS based on practice during the homework is what matters. Review the lectures notes and important high level concepts, and the rest of the class/tests will be based on writing MIPS code and doing different number conversions, arithmetic, machine code translation, etc.

Intimidation:
Rohr seems really scary, but he is a nice guy when you talk to him. If you have problems, go to office hours and he will help you. He realizes the class is really hard, so is willing to help students and do things like give a time extensions on 1 or 2 of the projects. He asks students for feedback too and truly wants people to do well and learn. Do not let his tough persona intimidate you. While he will not make any exceptions or go easy on any one student, he seems to be trying.

Exams:
Exam 1 was cake. Basic architecture questions and computer arithmetic. Review the lectures that will be covered and do the practice arithmetic problems and number conversions until you can do them without hesitating. Better yet, find a copy of a test from previous quarters to get a feel for what to know. I found a test from 1 year earlier, and the type of architecture questions were similar.

Exam 2 was reasonable, but everyone ran out of time. I thought I bombed it, but was saved by the curve. The material was very well represented by what Rohr said, but time was an issue. As such, practice number conversions a lot so you do not waste too much time doing them on Exam 2.

The final was also reasonable. A summary of the whole class. Review the important points from previous lectures and exams and you will be fine. It is open book, so knowing in which lecture to find a given answer really helps.

Specs:
A big concern is ambiguity of the homework specs. Unfortunately this is true, but perhaps exaggerated by previous posts. I found that I spent 10-20% of each homework either trying to understand the spec, or rewriting my program since I misunderstood something. Project 5 was the only project where that number was closer to 50% of my time spent due to ambiguity.

Projects:
Projects are tough -- what do you expect from assembly? Projects 1, 2 and 3 are C++/C projects that are relatively tough, but not too bad if you got through CS31 and CS32. Project 4 is MIPS. The project itself is not too bad, but will take a while to get used to assembly. I found it frustrating at first, but once I got used to assembly it was pretty straightforward. Projects 1-4 took me around 10-20 hours each.

Project 5 is killer. This is what makes everyone want to drop. I spent over 50 hours on it (30 hours of which was due to re-implementing parts of the ambiguous specs), and still got only 45/60 in correctness. There are so many little details that you just want to cry. The good news is that it kills everyone. I felt terrible only getting 45/60 after investing over 50 hours, but my TA said that most people didn't even finish it. If you turn in a complete and mostly working project you will be in the top of the class. If you can not finish it, do what you can and you will not be worse off than anyone else.

Project 6 is also bad. Another 50 hours project, but much better than Project 5. Project 5 is tedious, whereas Project 6 is long in that you have to implement a lot of straight-forward functions. If you like Computer Science, it is not too bad. If you are only a CS major cause your parents forced you, then it will suck.

Projects 7 and 8 are not too bad. Each took me 10-20 hours of line-by-line translating of C++ to MIPS. Just remember to test them thoroughly and you will be fine.

Difficulty:
Yes the course is difficult. What assembly class isn't? Yes Rohr makes the class even harder than it should be, but what weeder class isn't like that? It is hard, but not nearly as scary as everyone else portays. Work hard and you will do okay. I would suspect most people put in average effort and got a B or C. I put in 20-50 hours a week and got an A. This class is tough and will almost break you, but like every other class, if you tough it out than the curve will save you.

There are a lot of negative reviews for Professor Rohr, but I do not think they are entirely merited. I had him for CS33 in Fall 07, and I thought both he and the course are fine. (I will admit, coming into the class after reading the reviews, I was scared for my life.) Yes, the projects are very time-consuming (what do you expect from assembly language programming?), but they are also very straightforward. In fact, I've heard CS33 is conceptually easier than CS32, although I cannot confirm this, having never taken CS32. His midterms are also very straightforward and fair (the first midterm was, quite honestly, a piece of cake -- just remember your basic computer architecture material and how to do number conversions -- there's no programming!), and there are no "curve-ball" type questions that you may find from other professors. Professor Rohr started off each lecture with a "lecture challenge question" that reviewed the material from the previous lecture, and I found these to be very helpful. I found him very helpful during office hours (which I had to attend quite often... I think I'm slower than the rest of the students in that class) and was more than willing to work with me to figure out a solution to a problem. Prof. Rohr also extended the due date for some of our homeworks due to popular demand. There was a paper for the class, basically to discuss how you approached 2 of the projects, and the differences between programming in C++ and MIPS. This was another way to get free points. So, was this class a lot of work? Yes indeed. Was it hard? Fairly, although more time-consuming than difficult. Is the class required? If you are (un)fortunate enough to be a computer science major. Is Professor Rohr the mean, unfair, and incompetent bloke as described in many of the other reviews? My opinion is NO.