Avoiding committing many image files: Because different publications were using different versions of different figures, and I had many images that I could simply generate with my graph-builder, I just created a symbolic link to the directory where all of my images were being built to on my lab machine in the /images directory.

Style Issues: as I was writing my thesis, I learned that I am terrible for a few things

1. Putting Random Capitals in my sentences,
2. Not putting spaces before calling a \cite directive, and
3. Using " " rather than `` '' (the latex preferred/consistent quote characters).

Representing minutes: Representing individual minutes on a clock with only 12 hand-positions was a fun challenge. I ended up solving it by creating a modifier from the current minute and using it to modify the colour of the minute-hand.

Handling Hand-Crossover: Because I couldn't lean on hand length to handle the scenario when hands occupy the same value (i.e. 10:50, 3:15), I decided to have a location occupied by multiple hands become the average RGB value of each of the hands. This also looks super cool when paired with the modified minute hand effect.

Configurable Google Charts: I wanted to be able to select which of the metrics I would use to build my google charts (temperature, humidity, wind-speed, etc.) without having to manage some array of booleans. So I used this trick I learned at Delco where set each datapoint as an enum value that is a power of 2, then I set each checkbox value for the ones I would like to display as their corresponding power of 2, combining them with binary OR and then checking each the flags set on the result when constructing my chart.

TODO Fix Instrument Precision: the accuracy of the floating points displayed on the screen are far greater than what I am measuring them with... easy fix.

TODO Hook-up anemometer: I have a fan that should be able to generate an analog value for wind speed. It would need to be calibrated of course, but we can't really display wind speed / direction truthfully without it.

OS Independence: Because my original solution relied on a python GPIO library that wouldn't work for ubuntu server, I dug around a bit and found this dev.to post which revealed that in a very true-to-linux fashion, GPIO on raspi's is simply a series of file I/O operations. The first part of this challenge involved converting the whole thing to use c over python (a more appropriate language for system-level coding), the second part was pretty simple - control I/O to the appropriate GPIO files. It is a simple problem, so I guess it has simple challenges.

TODO apt-get repo: The next part of the challenge is to make this solution avialable via apt-get, It is annoying to have to pull the repo and make the project every time one of my pi's gets reset. Running the makefile is seldom a users first choice.