[an error occurred while processing this directive]
School of Computer Science & Engineering
University of New South Wales
Advanced Operating Systems
COMP9242 2012/S2
Lecturer Response to myExperience Survey
Overall
Thanks for the great scores! Looks like people enjoyed the course and
appreciated the learning (despite the pain!)
Free-form comments
We love reading all those comments on the best things about the
course. They line up well with what we're aiming to achieve, and
it's great to see students appreciating them.
Having said that, we're also grateful for the critical comments,
they help us keeping the quality high. I'm not going to address
every single one, but will comment on the common themes.
- More/longer lab sessions
- We did schedule more than twice as many lab slots as last year,
and did not receive feedback from demonstrators about a lack of
labs during
the term. However, this comment was made repeatedly, so we
explicitly followed up with tutors.
The ones we talked to did not experience overfull labs. However,
many students would walk in towards the end of the lab period and
expected to get marked, and the tutors ran out of time. We had
very specifically directed the tutors to not go overtime,
but close the shop at the end of the hour. It would really be
unfair to expect them to do otherwise.
So, based on that feedback, we can only conclude that the lab
times were sufficient, but some students had unreasonable
expectations.
- Video walkthrough of milestones
-
This is one of the places where COMP9242 is
“advanced”: we place great emphasis on students
working things out for themselves.
- Draft final report, more warning
- See above. The due dates are well known since the beginning of
the course.
TODO for next year: Add warning that the report
is significant work.
- Reduce M7 penalty
- While we sympathise with this request, we arguably already bend
the university's rules on late penalties in favour of late
submissions. We can't go any further than that, sorry.
- Requiring Posix is sad
- Philosophically we fully agree!
But Posix, for all its failures, at least gives you clear
guidance. Doing a proper cap-based OS would make the project even
more challenging, so we don't think that's an option.
Having said that, we do have such a project, and
it's very researchy. If you're interested, come
back to do your thesis on this project!
- Present alternative OS designs
- Fair point, would have to push out some other material instead.
TODO for next year: Consider adopting this
suggestion.
- Codebase issues, local setup
- We thought we ironed (almost) all of them out.
There is a
narrow set of tool versions which work for all the various bits
involved, some of the older than what is in standard distros.
Ideally things would work with recent toolchain versions, but
reality is different. We installed the most recent compliant
version of gcc in the class account and instructed students to use
those. It seems that many students instead used the default
version that comes with their distro, and ran into problems.
This isn't easy to solve, and the best we can do is to tell people
to follow instructions...
- How interact with seL4?
- Hmm, this is pretty much the content of the Week-1 lectures, and
the ample code examples (both in lectures as well as the supplied
code), plus the help sessions. We're unsure what's missing here,
more detail would help.
- Specs vague, show-stoppers unclear
- A degree of vagueness is intentional, to give you freedom with
the design, and encourage independence (as opposed to
hand-holding). However, it is clear that a significant number of
demos were show-stopped, and we should avoid that.
TODO: Revisit show-stoppers and provide
preventative spec improvements.
Non-issues
Looking back at the issues raised last year, it is good to see that
there is almost zero overlap with this year's list. This means
that we seem to have successfully addressed the following issues
raised last year:
- Ed forum dead
- Specific issues with some milestone specs
- Inconsistent feedback/assessment by lab demonstrators
- Odroid reliability
The one exception is on lab times and availability of help, despite
keeping our eyes on this. See above for our comments.
Final
Thanks for the feedback, and your participation in the course!
Gernot
Last modified:
09 Jan 2025.
[an error occurred while processing this directive]