At the end of October I started on a sprint to regenerate and rework all the game artwork.
Partly to upgrade to Retina resolution and partly to match the character art more closely to the user interface style. Less retro pixels & more children’s book illustration.
I estimated about 100 items of work that would take a week if I did 12 items a day. It took nearly 5 weeks. Doh!
What went wrong?
Disorganized source files. My asset organization system was a mess. I’d started out with neat nested folders but as the game evolved, the original categories made less and less sense. If I needed to do some exploratory art development, I’d put the files in a temporary location, then pull some of those into the game as a concept gelled. Some art needed extensive batch processing steps, like rendered animations. There were photos, scans, Illustrator paths, 3D files in several formats, render files and presets. Even so, it just got worse over the months and I knew I’d eventually have to overhaul the whole system. Eventually became now. Luckily I’m a digital packrat and rarely delete intermediate files.
Rusty illustration skills. What can I say, all the money I’m living on while I write THE OCTOBER GAME came from digital retouching, graphic design and 3d rendering. Yet I’ve let my hand-drawing skills deteriorate in a quite shameful fashion. It took many, many hours of painful scribblings before I got something I could live with on screen. I forgot how long it took to warm up the drawing muscles. I also forgot that if you’re not working from reference, you’re wasting time.
Work-like activity. In an attempt to stave off burnout, I sometimes took a few hours to practice UI design, prototyping and programming with little side-projects. It might be a game mechanic or maybe implementing a particle system. The things I learned weren’t a waste of time per se but I could only stand to be on the computer for so long each day and these projects burned up that time. These activities feel like work but they shouldn’t displace work. I’ve resolved to do game tasks before any other kind, going forward.
Low overall efficiency. Besides getting 60% as much done as I would predict, I typically underestimated complex tasks by 4X! The result? An exhausting number of things to do and a demoralizingly long list that grew over time instead of shrank.
What went right?
Getting better at planning. I’ve made great strides with the maxim “Measure twice, cut once” and “Third time’s the charm”. Before writing code, I write the problem down and I come up with at least three ideas. I break the work down until I can clearly visualize each step, then I try to predict how long it’ll take. I’m getting better at it each month.
Logging everything. You can’t improve what you can’t measure. Each and every day since July 6th, 2010 I move the cursor down to a line that reads “eliolhan work done – “. If I’ve moved my business forward, I put the information there. If not, I type “none” and I resolve to do better the next day.
Increased software writing skills. The code I’m producing now is significantly shorter, faster and has fewer bugs than work I did six months ago. Bug fixing now takes hours where it used to take days. I don’t throw away all that old code, though. I refactor and rework it.
What I learned?
Clearly I have to plan better, estimate better and work smarter. I’ve started by tracking efficiency on a daily, rather than monthly basis. I’m also tracking more fine-grained time estimates to build up a better baseline for later analysis. In the last ten days I’ve done as much work as the previous month, by keeping a closer eye on how I spend my time in front of the computer. We’ll see if it’s sustainable.
Next steps
I’m currently extracting the core classes in the game and placing them in a fresh app template. Along the way I’m doing some cleaning up and lots of bug fixing of older code.