[COFF] [TUHS] Interview question

[Adam Thornton]

> On Jan 2, 2023, at 7:58 PM, Larry McVoy <lm at mcvoy.com> wrote:
> 
> On Mon, Jan 02, 2023 at 02:13:45PM -0700, Adam Thornton wrote:
>> 
>> 
>> Which of these, if any, do you count?
> 
> Any of them that are entirely done by you.

Well, then, all of the ones I listed, if you'll spot me the Fellowship Of The Ring Atari 2600 cartridge before Thomas Jentzsch got me a little more display space, even though the cartridge you can buy (and the sources available online) are the later version.

None of those are world-changing, and, for instance, I'm a lot more proud of the current state of JupyterHub Kubespawner, which wasn't mine to start with, but first I added per-user namespaces and upstreamed that, and then I changed it from threads to coroutines, and upstreamed _that_.  The second change has made it a lot more reliable for everyone (the first was useful for Rubin Observatory and people doing things like we do, but not generally useful for JupyterHub under Kubernetes).  But both of those were building on (a lot of) existing work and were also made a whole lot better by review commentary from my team.  Special thanks to Russ Allbery who was quoted on TUHS a couple days ago, who really, really helped with the thread->coroutine conversion.

Sure, I can do decent work on my own, but other people helping me out improves it immensely.  I know that I tend to: a) under-test, b) write huge functions and source files when I should be doing things with more granularity, c) write like it's 1979 and keystrokes are precious and use nondescriptive variable names which make things hard to read.  Review of my work helps with all of these.

Adam