What I did in 2018

From Red Blob Games's Blog

My plan, from a year ago, was:

1. Produce explanations and demos from my procedural map generator experiments.
2. Go back to old articles and update them.
3. Work other developers on algorithms and articles.
4. Write a tutorial on coordinate systems and cameras.
5. Write tutorials about making interactive tutorials.
6. Become faster at writing explanations.

What did I do in 2018? I think I did reasonably well with goals 1, 2, and 6, and not so well on goals 3, 4, and 5.

Goal 1 - maps

• I improved mapgen2 with:
  • Smooth colors using gradients, an alternative to the discrete colors per biome.
  • Icons for mountains, forests, grasslands, water.
  • Slider to control jagged vs smooth coastline.
  • Sliders to control temperature, so that you can have cold poles / warm equator.
• I improved the home page by adding links to many more pages.
• I built mapgen4, a map generator that lets you draw your own constraints and generates the rest in real time.

I'm really proud of this work. It brings together the things I learned last year during my map generator experiments and puts them into something fun and cool.

Goal 2 - updating existing pages

• I simplified the C++ code on my A* Implementation page to make it easier to read.
• I added support for touch events to my A* Theory page and other pathfinding pages.
• I added another diagram to my Tower Defense page.
• I improved my Visibility page by (a) numbering the points in the scan, (b) numbering the triangles in the output, (c) making the draggable elements respond on mouseover, (d) simplifying the underlying implementation.
• I updated my Hexagon Implementation code generator to make Java, JavaScript, TypeScript, C# output more closely follow standard style.
• I reimplemented my Hexagonal Grids page, added many more diagrams, and made it load faster; see writeup part 1, part 2, part 3.
• I updated 92 pages to work better on mobile and tablet.
• I improved the cache behavior to make pages load faster.
• I made the pages print better.

I'm quite happy with this work. It's not highly visible like a new project is, but one way written web pages are different from books, academic papers, or videos is that I can continue to improve them over many years, even decades.

Goal 3 - work with others

• This year I wasn't a mentor for Google Summer of Code, but I did regularly chat with this year's mentor and student, who produced the interactive version of Chapter 5 of Artificial Intelligence: A Modern Approach.
• I worked with Vladimir Agafonkin on a new guide to Delaunator, his ultra-fast Delaunay+Voronoi library. I had been using this library for a year, and felt like I had learned its secrets. The library documentation didn't cover most of what I had learned, and I wanted to share them so that others wouldn't have to rediscover everything the hard way.
• I went to Game Developers Conference and Roguelike Celebration, where I met lots of game developers and got to chat about the algorithms that they need to use in their games.
• I'm on the Roguelike and ProcJam Discords, several Slacks, several subreddits, and StackExchange.

I've come to the conclusion that directly working with other people on code isn't a good match for my skills. I'm a reasonable coder, but my coding style and work habits are at odds with what most people need. Instead, I should focus on my writing. Talking to other developers about what algorithms they use and what kinds of problems they need to solve will help me decide what topics I write about in 2019.

Goal 4 - coordinate system tutorial

• Mapgen4 took a lot of my time this year, and I didn't attempt the coordinates tutorial. I think that's ok. In terms of meeting my goals, it's a “failure”. But I don't consider working on mapgen4 instead of coordinate transforms to be a “failure” of the work. Instead, it's a failure of how I set the goal. Maybe instead of listing a specific tutorial in my annual goals, I should list the type of work I want to do.

Goal 5 - metatutorials

I had hoped to write up a lot more of what I've learned, but I only wrote a little bit:

• I wrote about the overall structure of my pages, including some of the things I've struggled with the most, and how to solve them.
• I discovered a way to automatically redraw a <canvas> when the data it depends on changes. I'm now using it for several of my pages to improve performance and reduce errors. I also adapted it to work with D3.js.
• I wrote how about make "scrubbable" numbers in ObservableHQ.

Goal 6 - become faster

I worked on lots of tiny (<1 day) projects, often to learn new tools, techniques, and libraries that might help me work more effectively.

• I continued playing with ObservableHQ, with pages like Centroid and Voronoi polygons and Reshaping Distributions. It's a really cool system, but I ultimately decided it's not for me.
• I continued playing with Vue.js, with pages like Hex half directions and Hex rectangular wraparound and tree render. I rewrote the Hexagonal Grids guide in Vue, and it was a great experience.
• I improved my private mouse+touch handling library, Draggable.js. I had first developed this last year, as an alternative to d3-drag for projects where I don't want to pull in D3.js, but over time, I've found myself using it more and more, even when I am using D3 already.

Some other short (<1 week) projects:

• Modifying spelling
• Procedural face generator
• Math explorables
• Voronoi on a sphere
• Planet generation

Next year

The shutdown of Google+, the censorship of Tumblr, and the deletion of photos on Flickr made me think about social media and longevity. I'd like the things I spend time on to last a long time. It's worth spending more effort on something if I know it's going to last decades. I've been increasing the amount of time I spend writing to my blog (RSS feed), and trying to avoid content that's I post only to Twitter or other platforms. You can see this in the blog's sidebar: I had averaged 10 posts per year in the previous 10 years, and wrote 37 posts this year alone. I want to focus on the web and not on proprietary platforms.

I had gotten stuck a few years ago, and experimented in 2016, 2017, 2018 to improve my skills. Here at the end of 2018, I'm pretty happy with where I am and what I've learned. My goal is not to write software or make interactive diagrams. My goal is to write explanations of math and algorithms. I write code and diagrams to support the explanations, not the other way around. In 2019 I'd like to focus on improving my explanations and writing new ones.

Labels: annual-review