Lukas @jemus42@scholar.social

I tried the Google maps AR walking navigation thing today. It is oddly futuristic, yet also somehow "why hasn't it always been like this this is so neat"

You ever need to do regression through the origin and really, like, _really_ want to cancel out parts of sums? Also, my google foo yields depressingly little info on the whole thing, at least on a level I can grasp (not familiar with regression in matrix notation yet, and multiple regression is also a topic for the coming weeks).

Sigh. Math is just too math for me.

Oh look it's exported/uploaded with background transparency. That doesn't help.

I need to derive the variance for the beta estimator for simple lineare regression through the origin (RTO), and let's say my math foo is insufficient.

I've been spending more time than I'm prepared to admit on the question of which fonts to use for the .tex template for… weekly assignments that nobody really cares about with regards to typographic sophistication and aren't even required to be TeX'd at all.

I switched my .tex template to scrartcl and now \emph doesn't do anything anymore and I'm not even sure where to start trying to fix it (╯°□°)╯︵ ┻━┻

If you're into self-inflicted pain like me (or just want to try to undertand how to kind of manage a TeX-distribution and use TinyTeX instead of TeXLive et al) … uh, long story short: This tool: https://ctan.org/pkg/texliveonfly can parse your .tex file and auto-install all the packages you'll need. It's magic. I love it.

Turns out my package build on travis is failing because apparently the check process uses http://worldclockapi.com/api/json/utc/now to check for the local time. Which is currently 403. Which is… a bummer? One might assume that "figuring out what time it is" would be a solved problem, but alas, 'twas naive.

Ha! Finally found the issue. My best guess is that the cached data at cloudflare differs between my location and the location of the travis servers, otherwise I can't explay slightly different API response content for the same query at the same time that triggered the bug.

MFW a test repeatedly fails on travis but I can't reproduce it locally or on a different machine

Sigh. I wish all the #rstats Twitter people were on here. I really don't like having to open Twitter to see what's new these days.

I love it when I think a problem into the internet and the internet thinks back and solves a problem for me.

I'd really like to validate a URL by making a testing-GET request, see if it returns a status code 200, and *then* I'd like to make the *sincere* GET request where I may retrieve more than a megabyte of data. I feel like I have the option between "guessing a URL works and possibly failling" or "hammer the API and waste bandwidth".

I might not understand how HTTP works, that's also an option. Or I just should be better at failing in code? Idunno. Things are hard.