[paxys]

In my experience it's always the former. Back in university the surest sign of a good class was that you were allowed to "cheat". The professor encouraged using any and all outside materials for your assignments. You were allowed to work with other students on projects. Exams were all open book. You could use open source code. Such classes always led to the best outcomes for students who were truly interested and engaged, and those who weren't had no way of faking it. On the other hand classes where professors spent more effort on detecting cheating than teaching the material were always duds.

[danpalmer]

Huge +1 to this. I had one lecturer who would essentially rant at us about security vulnerabilities every week for a semester, and then write a paper based on his rants.

Some students ignored his rants because they didn't find the subject interesting, did the past papers, and then got to the exam and did terribly, then complained about the fact that "it's all based on his rants".

Others found the ranting engaging because it was deep dives on obscure bits of computer security history. My year he spent ~4 weeks going on about Stuxnet, including deep dives into the wider political context. When I got to the paper, one of the few questions we could choose, for 50% of the paper, was just "What was Stuxnet". I wrote pages and pages. Figured out from marking that I got full marks on that one. I always did great in that lecturer's modules despite never taking notes and rarely doing any targeted reading.

To show the level of engagement, I once turned up to one of these lectures 10 mins late, with most of the class already there, and the lecturer said "oh I wondered where you guys were, I guess I'll start again". He knew we were the only ones who cared about the class.

[sitkack]

I would be on the edge of my seat for all four hours of Stuxnet! The Musical. Sounded like a really fun class.

[danpalmer]

Even better was the second year of the class (this lecturer lead the security side of the CS degrees). I took it in autumn 2013, just after the summer of Snowden. The class was just week after week of "I TOLD YOU SO! Look at these leaks! Look at these NSA details! Look at what we happens if we take this bit and stick it into LinkedIn and Google Street View and see where the developer of XKeyscore lives! Look at where this undersea cable goes, isn't that a bit close to Diego Garcia?! Look at this submarine and the bit carved out for an undersea cable splicing facility! I'VE BEEN SAYING THIS ALL FOR YEARS!"

As much as it sounded like the rantings of a tin-foil hat wearing madman, he was in fact quite accurate through a combination of having worked in government in computer security at a fairly high level for a while, and having deep technical knowledge about what was possible.

I miss his lectures, they were excellent, but I hear he's still going strong scaring freshers and running the on-campus teaching union, which is nice to hear.

[sitkack]

I love playing BoE (back of envelope) Gedanken Experiments. Splicing a undersea cable or recording and analyzing all the SMS traffic in the world seems a little more tractable after you do the math.