Micro-projects - A Commitment Mechanism

There was this brief period of my life where I got into life-drawing. The first time I went to a class, I was terrified, because I straight up didn’t know how to draw, despite having had a nebulous goal of “get better at drawing” for several years. Yes, yes, it’s counter-productive to feel embarrassed at your abilities in a class designed to improve them. And yes, yes, no-one actually cares or is paying attention. But my body doesn’t know that instinctively, so I was intimidated.

Before I knew it, there as a naked person in the middle of the room, and the teacher had said “alright, let’s start with 30 second poses as a warmup”. It’s genuinely not hyperbole to say that the next 30 seconds were life-changing. Because here’s the thing: you’d have to be Da Vinci to draw a masterpiece in 30 seconds. And if you’re trying to get anything out on the page in 30 seconds, you need to hustle. That sense of urgency completely released me of any sense of self-consciousness, perfectionism, any thought other than “draw”. And for the next year or so, while I went to those classes, I experienced a rate of improvement like never before.

Lately, I’ve been feeling stuck. There are so many interesting things to learn about, investigate and build. I - like most engineers, I imagine - have a note on my phone called “Project Ideas” that grows ever longer. It’s so easy to have ideas (and it feels great too!), but it’s far harder to execute on them.

The behaviour that I typically exhibit is:

Have a new project idea and feel really excited by it
Make a start, lay out foundations
Tie myself in knots trying to make sure I do it perfectly
Get distracted by something else because it’s not fun anymore

So I’ve decided to adopt one of my software development policies as a policy in my own personal development: “assume failures will happen and respond accordingly, rather than trying to prevent them entirely”.

The goal of these projects is and always has been a) learning something new and b) having fun. So if I typically stop having fun after a few weeks working on a project, then I guess it makes sense to no longer try and work on projects for longer than a few weeks.

These are the rules:

Projects last 3 weeks - no more, no less. Pick an end date at project inception.
Whatever state the project is in on that date, it’s done - stop working on it.
Before starting, scope the project ruthlessly so it’s achievable in that period of time.

To be able to scope projects well, I’ll need enough background knowledge to be able to make an intelligent judgment call. So perhaps for a week or so before starting a project, I’ll aim to read docs, read background info, and just generally immerse myself in supplementary materials, without any pressure to be executing.

In the same way that the 30 second drawing warmups removed the pressure of perfection in the life-drawing class, I’m hoping that these rules will force me to make something finished and imperfect in these micro-projects.

Some ideas that are currently floating around:

A Monte Carlo path-tracing renderer
Something to do with stylized material shaders
Something with a purely functional language
Something with a Lisp
A FVM fluid simulation of some kind
A SAT solver
Something with an FPGA (or simulation thereof)
A Kalman filter in C
A photogrammetry tool
A non-linear optimizer

We’ll see where we end up! The aim is to expose myself to as many new topics, paradigms and ideas as possible - to go wide before I elect to go deep on anything.

Project data will end up here, and I’ll chuck up a post at the conclusion of each one.

See you at the end of project 1!