Comments on our Bespoke Survey

I'm very pleased to see that all but two of the enrolled students completed the survey (bribing with a t-shirt may have helped ;-) This certainly provides valuable feedback.

General comments

The results mostly speak for themselves. The course seems to be in good shape, overall satisfaction remains very high, some critical comments notwithstanding.

We are grateful for all the positive and encouraging comments, I won't discuss them, but be assured they are well received by the team! Let's look at the critical comments that address issues under our control (and changing which would not obviously be at odds with the overall objectives for the course). Also, I'm skipping comments that are already addressed in the myExperience response.

Blue font indicates concrete action items.

Q1: Quick evaluation

It's great to see the strong approval scores for all people involved (individual lecturers as well as the class of tutors), with "excellent" clearly dominating. It's pleasing to see that their commitment is appreciated.

Q5: Content Balance

The vast majority thinks the balance is right, and the too-much/too-little also are pretty balanced, so it looks good.

Q7: Worst Things

Workload, forum and lab load came up in the myExperience survey and are addressed in that context. Others are discussed below.

Tutors being unprofessional?

This is a single complaint, more than balanced by many strongly positive comments about the tutors. I see no evidence that this is a real issue.

Limited relevance of lectures to the project after Week 3

Well, the course is Advanced Operating Systems, and it's more than just the project. It aims to provide more depth in a number of directions we consider important (an obviously biased view, no apologies!) And the vast majority of students found all lectures relevant (Q11). Furthermore, answers to the question about the most useful material (Q13) covered pretty much everything, and there was few comments at all about what to exclude (Q15).
In summary, I see no reason to drop any particular topic.

Exam not enjoyable

There clearly are widespread opinions about this, some people like it, some don't. I like it for the exact reason several students stated: it's a different way to assess, which is a match for the different kind of course AOS is. It really tests understanding, rather than rote learning.

More practice for exam, point to papers students expect to have read and discuss in class

This is related to the previous pont. I've tried that in the past, but almost no-one reads the papers as long as there are no marks involved, so it's pointless. But we do list papers for almost every lecture, and critically reading those is the best preparation for the exam (beyond actively attending lectures).
I can only repeat: Actively participating in lectures, including asking and answering questions, is the best way to acquire the insights the exam tests. The second best preparation is reading the papers listed.

Odroid reliability

While this was clearly less of an issue than last year (following improvements we had made to the infrastructure), the comment still came up multiple times, and we're taking it seriously.
ACTION: We've got one student of the 25T3 class employed on doing a significant upgrade of the cluster management infrastructure. Hopefully this will fix the problem.

Make M8 due Mon of W11?

I thought this would run foul of UNSW rules, but upon checking it seems we can get away with it.
ACTION: M8 deadline will move to 9am Monday of Week 11.

Windows and Mac support

This was a single complaint (while common the previous year), so I assume it's mostly under control, but we'll keep ensuring that the environment is well supported on non-Linux systems. There are plenty of people in the TS lab doing seL4 work on Macs, so the support is generally there, within the limits posed by CSE's level of support for non-Linux machines.

No multicore etc?

Do you really think the project should be even harder???

Q14: What should be added?

More on formal methods in the seL4 context

I used to do this in the past, but it was one of the less popular topics, so I dropped it.

Asynchronous APIs like epoll, io_uring, kqueue

Good suggestion, and fits well with coverage of LionsOS.
ACTION: Cover Linux asynchronous I/O together with LionsOS.

More on other microkernels and their implementation

I'm not really the expert on other microkernels. But they all have in common that they are slower, less micro, and (except for some research toys) not verified, so what's the point?
We had some more detailled coverage of Mach (and its many deficiencies) in the past, but it seemed too much like bragging. But some of it is on the EOS Microkernel lecture, although the coverage isn't that extensive.

Unikernels

They are really a fancy name for library OSes (there's a whole lecture in EOS on this). But I guess mentioning the library OS concept is reasonable.
ACTION: Mention library OSes.

Q15: What should be scaled back?

Only a few comments here, and generally about things others really liked.

I take the point re the Apple lecture, I was annoyed too.

Q21, Q22: Quality

Looks it's generally pretty good.

Q24: GDB

A number of comments about GDB breakage and poor support for multi-tasking. Part of this may be documentation rather than software.

ACTION: We'll see what we can do, after the cluster infrastructure work is done.

Q26–29: Partner Issues

While to a degree unavoidable, and less of an issue than the previous year, it remains a significant issue.

ACTION: We'll try to provide more support for partnering – see the startup checklist!

Final

Thanks for the feedback, and your participation in the course!

Gernot