Advent of Code
It's around this time of year that my colleagues and the development community as a whole start going on about Advent of Code and, every year, I have to explain that I just don't see the point.
It's around this time of year that my colleagues and the development community as a whole start going on about Advent of Code and, every year, I have to explain that I just don't see the point.
The whole idea of Advent of Code, and coding katas in general, is that you're given a "toy" problem to solve and you want to learn Rust or you want to try out TDD (you should try out TDD by the way!) So you try and solve the "toy" problems using Rust or TDD or whatever you're grasping at.
But, at the end of it, what do you have?
Fine, you've got a better grasp of Rust or TDD and that's really great... but what value can you gain from the actual software you've written?
I've got "a thousand and one" pieces of real software that I actually "need" in my life.
I air quote "need" because many of these projects are quite pointless hobby projects that add no real value to the world, I just want them.
But not always entirely pointless... without my Lume based, static site generator, you wouldn't be reading this guff right now.
But for one of my more trivial examples, I've been working for ages on a project to connect an obsolete GPO-746 handset to modern mobile networks just to have a retro-hipster telephone that I can play with and show off to my friends.
For this project, I found that I needed an Android App. So here I am learning Kotlin, and finding that I need to write multi-threaded code. And then I want to easily test my hardware, so I need a native Linux testing application that shares code with my Android App and I find myself doing native Kotlin and using it's foreign function interface to talk to a Linux USB library.
Multiple learning opportunities flood out of this project and at the end of it I've got a working application that I want, not just a solution to a puzzle that doesn't actually do anything.
My telephone project has an AVR microcontroller in it and I like to program AVRs in assembly language.
For my AVR projects, I was using GAVRASM.
But you'll notice that link is to The Wayback Machine... it seems my beloved assembler has "died". I have a sad suspicion that it's author, Gerhard Schmidt has gone to the big morning standup in the sky.
GAVRASM didn't have a fully Open Source compatible license and I want a replacement that does. The available options don't quite "float my boat" and so I'm now working on my own replacement.
What's more, just to troll the "Macho Programming Bros" just a little bit more, I'm doing it in Typescript!
These are just a couple of examples, I haven't mentioned Zig, real time audio, Python, video processing and all those other things on my mile long TODO list.
But this is my point. There are so many opportunities to learn that leave you with a fun or maybe even useful piece of software at the end.
Why use up your valuable coding time solving puzzles for the sake of solving puzzles when you could be building stuff?
I can't criticise the "Programmings Bros" for telling us that we're doing everything wrong... and then come along and do the very same thing.
If coding katas work for you, if you can't wait for December to come along so you can start this year's Advent of Code - then, please, ignore me and carry on.
There aren't any imps who are going to poke you in the nether regions with pitchforks for eternity if you don't follow my one true way.
Although... I'm still obviously right and you're obviously wrong. ;)
The original version of this article was just a rambling stream of conciousness and you don't want to read it... honestly you don't.
But it was inspired by a couple of excellent videos on Live Overflow:
CTFs are awesome and CTFs are terrible.
I'd still highly recommend these as a similar perspective on CTFs which are, more or less, the same thing as coding katas but for security people