DEV-DIARIES

Dev Diary 2018-12-09

I want to start getting into doing development diaries again, because I enjoyed reading the old ones I did back when I was still working on NiceBSD. I’ll try my best to make them useful and enjoyable for you as well, dear reader ;-)

I’ve just gotten word that I’ve been approved for a Pecha Kucha talk for the Winter Cycling Congress that will be happening in Calgary in February 2019. It’ll be along the lines of how running a fleet of simple, cheap-to-run bots and websites can create an enthusiastic (and data-driven) bike culture in cities.

Speaking of those bots, I really need to figure out how to get the override system to work better for the CoffeeOutside bot. The success rate for overrides is miserable, and the bot is a little too good at doing its job.

Work continues on the Cyclepalooza site, as we prepare it for Winterpalooza 2019.

I’ve just broken ground on yycbike.info, which will be an interactive infographic site along the lines of what Bret Victor has shown. Right now it’s just running the old cycletrack Tangle site from the pilot presentation, but I’ve got a lot more planned for this one, with the official “launch” happening at WCC.

I’ve moved most of my sites to Hugo, which has been kind of a slam dunk for writing new stuff. Live reload is magic stuff, and a real game-changer for my workflow. There’s two more sites left for Hugo-ization, and that’s daveops.net and vr.daveops.net. The VR one might seem like a weird inclusion, but I’ve got an idea of making a Hugo theme using A-Frame.

On the R+D front, I’m working on getting my LPI certification so that I can start teaching Linux for one of the local non-profits. It’s been a light and breezy affair so fair, but there’s been a few little nuggets of information that I hadn’t discovered over the years, so it’s been a fun experience.

I’ve also been playing a bunch with Kubernetes as of late. I’ll be honest, I’ve had a lot of problems with it, and there seems to be a documentation gap between “Hello World” and full-blown “hundreds of nodes running a gajillion containers.” I want to believe in the vision, but it seems like there’s a lot of hype and sales talk, and not enough on the problems with it, like an ops version of MongoDB. There needs to be more conversation on the gap between running Kubernetes on your laptop to running a few compute nodes securely in a production capacity. Maybe kubeadm going GA in 1.13 will help smooth out that path?

Advanced Kubernetes Documentation