Post mortem

What went right

• Prioritise - made sure systems were finished in the right order, so that other could be built upon them. For instance prototyping the car controller, then creating the town, then creating the ability to pick up passengers.
• Time management - Even though the blog records daily activities, it was usually only a few hours per day spent developing, sometime more, sometimes less. It was important to get chunks of work done in order to be prepared for the next day. It also meant I could think about the next steps, and how it'd get developed.
• Don't aim for perfection - everything could've been better. I tried to get it 80% done and all functionally working. I didn't spend time on efficiency or optimising anything. When time is scarce, quality tends to suffer. I tried to ensure the quality was lost in parts that wouldn't harm the final result. For example the car controller didn't have gears or sound correct, the acceleration could've been improved, but it was functional enough to be usable. The car npcs in the game aren't intelligent they just follow a path, it's not ideal, but it is functional enough. The passenger missions could be better, e.g. you can pick up Guy After Date, and then pick up Guy Going To Date, you also drop them off at random places rather than a a particular building. It's not ideal, but it is functional and serves it's purpose. All of these could be improved given time, but what was created works sufficiently  well for the game to function.
• The TTS voices were pretty good. I did records several voices and put them in, but the volume was too low as I didn't edit them and didn't realise until I played the game.
• The character image generation worked well. I did have to create several images and pick the best ones, but it was a lot quicker than thinking and designing them myself.

What could've been better

It's difficult to judge what "better" is, since given the dev time, I think everything could've been developed better. So "better" in this case is with consideration to the strict dev time and unpredictability of how much time would be spent developing on any one day.

• Project management - While the project was complete, I think next time I should've left 0.5 - 1 day to play test and add polish.  There were a few things that were rushed at the end.
• Sound management through the UI - the music was added late, as was some of the extra voices. This resulted in sounds being drowned out by the music. It would've been useful to have been able to mute the music or change the volumes from the front end, even just being able to mute the music would've been helpful.
• Audio balancing - The speech audio files should've been normalised and amplified as needed. Currently they can barely be heard over the music. On the last day I did use Audacity to improve the sound volume, but by the time I put it in the game, balanced the music and sound with a mixer and recompiled it was past the submission deadline. 
• Made a list of final improvements needed - I should've spent the last day play testing and then made a list of what final improvements / polish could be added. There wouldn't have been time to do everything on the list, but I could've then estimated each item and weighed it's benefit with the time it'd take to do the task.