The Dreamer postmortem (and timelapse!)
Posted by johnfn (twitter: @thedayturns)
December 20th, 2010 1:13 pm
So the Dare is over!
(It’s been over for a while, aye, but I was kinda dazed, and I started playing other people’s games for a while to restore myself.)
I think that this is where I mention the major problems that I had. But I really feel like I lucked out this time, because the creation went off without major hitches, and I even managed to adhere to a schedule (!!!)
Still, some notes:
- I’m still not exactly sure what kind of starter code we’re allowed to have. My starting code put me at about the same level as Flixel (but without map support), but that still disadvantages us to people who use things like game maker and get stuff like map support and room support built in. I don’t think I had all of that until halfway in!
- This was the best I’ve done with abstraction, I think (at least up until the last few hours where I was just cramming in fixes). It really paid off, too – the latter half of development was relatively breezy.
- I was insanely productive. This kind of game usually would take me around 2 weeks to make.
I’ll stop patting myself on the back now:
- My biggest problem by far was the lack of a good map editor. If you look at the timelapse, you might notice that it took me something ridiculous like an hour or so to develop a single map. Going back to make changes (of which I had to do several times) took a while too. Editing the raw data in Flash was atrocious because Flash just dies with that much text – I eventually switched to FlashDevelop, which is lightning fast. The good news out of all of this is that I have finally discovered how to solve this problem, and with it, another problem that I’ve wondered about for ages. Essentially, the tile editor must handle ALL of the output. My current setup was just outputing the raw map data, which I had to put into a class and hook up with exits to other rooms (which later led to nasty bugs). Editing maps and writing code should be completely separate activities. I really wish I could have made a more expansive world, and perhaps played around with the alternate world idea a little more. Ah well. I’ll be taking a few days to make my tile editor better (again)
- I never made a sprite for the main character and eventually just had a stupid smiling pixel. I need to learn how to do art better. I did some tilesetting a few months ago, but I never made characters. This needs to change.
- I didn’t have any animated things. I wanted to make the flowers move around, for instance, but I was preoccupied with other things, and worse, I wasn’t sure I knew how. Now, come to think about it, it’s not that hard, but it requires some more abstraction (yay!)
- The worst bugs in the game were related to collision detection and map transitions, two things that I should have (and could have) done before the Dare. Next time!
All in all, LD was a good experience. Makes me wish I could work at that efficiency all the time…