Survey 2019

I used survey monkey for this year's survey. Here is a PDF of the results.

Thanks again for all your feedback, it is much appreciated.

Here are my overall observations:

I had feared that Trimesters would destroy the essence of the course - challenging practical application of theory in the assignments. I'm very very happy the course ran without major hiccups mostly intact, with generally quite positive feedback. Phewww.

I tried the following changes, which seemed to mostly work. Though some further tuning is required and students did not really get started until week 2.

I'm releasing a the warmup before semester starts so students set up their environments prior to semester
The warmup has more wiki-based guidance on getting set up
I'm providing more guidance on using your own machine (which is more efficient than remote CSE usage.)
The warmup is less exploratory and more focused on simply learning to build and submit and assignment.
It's due earlier in the semester
I'm changing the order of the lectures to move concurrency earlier so in course so the assignment can be started earlier.
The first substantial assignment is done as an individual so can be started ASAP.
I've remove some of the components of the later assignments, and aimed to remove common pitfalls that were good for learning, but potentially time consuming to get out of.

The reduced number of tutorials were noted.

Part of the feedback was inability to cover all the questions - this was true in 12 (and 14) week semesters. The aim is to select based on demand/interest/need. I'll provide more guidance in 2020.

Piazza is generally viewed positively (suspect is is partially due to my responsiveness), though its web UI is becoming dated. The apps are clunky, but I'm okay with that, you don't have to use the apps.

The etiquette guide seems to have reduced the noise a little this year, I'll point it out more explicitly.

I've contacted piazza with some suggestion on search improvements. They seem to be responsive and have made some changes to improve the search functionality.

I'll provide some guidance on reducing Piazza's spammy-ness.

The switch to GIT was a success. I'll retain it and gitlab going forward. It is a little concerning that some students avoid it when collaborating. I'll provide some more encouragement to learn it, it is really the dominant industry revision control system, so worth learning independent of the course.

Lecture attendance is dropping to unsustainable levels - lecturing < 30 students in Clancy auditorium is like talking to myself. It also means I don't get good feedback from a wide range of the cohort with so few in the room (part of the reason I went too slow in 2019, despite removing the material from the course, and aiming for only 9 weeks in 10). I'll do business as usual in 2020 to get another year's experience with trimesters before changing anything major regarding lecture delivery.

It seems my videos are still positively viewed and are sufficient for most with roughly an even split between "lecture to fast" and "lecture to slow". I guess the downside is that I lose these students from the auditorium.

I'll see if Camtasia can be tweaked to have keyboard shortcuts which seems to be the most common feedback. I'll provide some guidance on downloading mp4s and the table of contents which some students seems to have missed being available.

The benefit of my videos over echo360 is reducing, so I'll keep doing them in 2020. I may make echo360 available in parallel.

GDB was commented on, I plan to provide a more systematic intro to debugging for the course as an optional side module.

ASST1 was probably too easy. ASST3 remains challenging, despite reducing the scope compared to previous years. I don't plan to change much for 2020.

It's a little concerning that about 20% of the cohort skipped the assignment prep tutorials.

The early bonus from past years to motivate starting early did not work as fairly under the tighter timelines (though they were not that much tighter). I'll look at how best to improve the fairness.

VScode was a surprising winner among editor choices, it positively dominated. It makes more sense to provide some more guidance on its use.

I'll switch to HTTPS for gitlab access to avoid UNSW firewall issues.