2024-03-07 Comments

[Robadob]

Had session with 4 PhD students today.

Encouraged them to provide their own feedback either here or via email.
One did note that they hadn't understood everything, but it's given them alot of ideas to take back to their code.

Definitely think more diagrams would help, e.g. to describe binary search. May want to even explain the interface to bisect_left() here as its very quickly glanced over.

---

I tried to keep track of timing (atleast through the first half)

Prof Intro: 15 + 10 mins
Function Prof: 20 + 10 + 10 mins
Line prof: 20 + 15 +10 mins (more difficulty here with package issues)
prof conclusion: 5 mins
break: 15 mins
optimisation intro: ~10 mins
data structure + numpy : ~1hr
lunch: 1hr
update py/latency: ~45 mins
opt conclusion: 5 mins

Definitely space for more content, whether that be going over obvious stuff, timeline profiling or optimisation exercises at the end.

[willfurnass]

One did note that they hadn't understood everything, but it's given them a lot of ideas to take back to their code.

I think that might be the outcome for others in future sessions too, but it's not necessarily a bad thing provided the material remains publicly accessible and people know which parts to revisit after the session and why they should care.

[Robadob]

I think that might be the outcome for others in future sessions too, but it's not necessarily a bad thing provided the material remains publicly accessible and people know which parts to revisit after the session and why they should care.

Oh, they were very positive about it. That wasn't framed as a problem. I'd been quite clear that they don't need to come out of it, for example understanding how to implement a list or set from first principles. That the explanation was to help ground the good vs bad patterns.