Synthesis is finished and it’s about time for a postmortem! Let’s do this!
There’s also a timelapse, if you’re interested.
What went right
I’m guilty of being a little pleased with what I’ve done during the weekend. Not in terms of a perfect game – rather in terms of my personal progress during the compo. I didn’t really like the theme, but I did manage to finish, learn a lot of new things and have a lot of fun.
- Picking Flashpunk as a library. Flashpunk is awesome. It’s very intuitive, quite well documented, and has many great features. I kinda prefer Flashpunk a little over Flixel (which I used in my last entry) because it feels more code-oriented, but I’m not that sure if this is a good thing – the best games I’ve seen so far were done in Game Maker, Multimedia Fusion, AGS, Stencyl, where the workflow seems more assets-oriented (though I admit I’ve never used any of these – probably should give them a try).
- Community help. Flashpunk tutorials (mostly those by Zachary Lewis, examples from Chevy Ray’s keynote) helped me remember stuff, and I even managed to find some preloader example on Flaspunk forums, as well as a nice Kongregate API wrapper class. Wow, Kongregate! I’ve never done that before! The community help was crucial here. Thanks to all who helped me!
- Physics. I aimed for asteroids-like movement with some friction. Attraction and repulsion mechanics seemed fairly complicated, but I spent a lot of time tweaking it, so it’s probably the best I could do with area-based meachanics. Interatomic attraction works quite fine and is even flexible enough to modify in possible future versions. I feel controls may still be too hard for many people, though – just like in my other games.
- Graphics. From the start I decided on something very simple and that was a very good idea. I did waste some time on placeholder graphics I didn’t want to use in the final game – see “things to consider” section. I think the text bubble thing – done in last few hours of development – bears a really nice effect, and I’m really glad I decided to add it. Also, particles! Wow, I did particles! I’ve never done them before!
- Sound and music. Wow, autotracker-bu is SO, SO, SO awesome and easy to use! I love that there IS freaking music in my game (without the generator tool, I couldn’t make it if my life depended on it)!
- Food. You can see it in my journal. During saturday I ate many good things – mostly because I didn’t have to prepare them myself. On sunday I just ordered pizza.
What went wrong
Besides being so horribly tired after the compo, I don’t feel that I did anything different from the last time, and this seems to be my main problem – dealing with the same issues over and over. Damn, I must try harder to do something about these next time.
- Performance. There’s a reason for having only about 20-30 atoms at most on the screen – with about 70, the game becomes unplayable due to computing too many atomic interactions. There some space for optimisation here, but I didn’t have time to think about it. Now I can just hope it works on every older machine.
- One level, hi-score based gameplay. Yeah, that’s cheap. Next time I will try to make something with a story an more than one level.
- No programming preparation. Picking AS3 without any preparation. Dammit, it’s the same mistake I did before! I don’t write in AS3 for other occasions than Ludum Dare (mostly C/C++), so it’s very painful to try and remember how that cursed variable scope works in this goddamn abomination of language!
- Working without any pre-compo framework. I didn’t have anything prepared so my code looks awful – again. On the other hand, preparing too much code before the compo could mean that I’m constrained to a specific type of game… However, I feel I should do this at least once – I wonder if preparing some initial framework would make my games any better.
- No plan, no schedule. This one went actually worse than the last time, but partial knowledge of Flashpunk did save me this time. If I were to learn everything from the beginning (like I sometimes do), I would fail miserably. I think I might try to prepare some easy to edit list of “todo” things, to maintain priorities when developing a game. This time I was lucky, I guess.
- Rusty chemistry knowledge. I had to look up many chemical compounds on the web – especially their English names. I missed a few simple ones, and many complicated ones. This is also connected to the point below.
- Theme. After some brainstorming, I got some general idea for the mechanics, but during the development, it turned out that I was wrong about how chemical bonds work and had to think about some other solution. I decided on a list of possible molecules – there are 45 of them in the compo version (I will probably make another post about this).
- No coffee on Saturday and I didn’t expect that. WTF? Oh well, I had to drink tea.
- Too much coffee on Sunday. Too little sleep. I had to rest for another day just to get into some kind of shape.
Things to consider
- Placeholder graphics. I find that using simple squares isn’t enough. At least static, red or green squares, without any animations. If you’re aiming for entities with animation, you should prepare some simple graphics beforehand.
- If you think that using some game making software is lame, go and look it up. Being a programmer myself, I find that creating a polished game in 48 hours from scratch (even using a wonderful library like Flashpunk) is really hard. Not to say that using such software makes this process easier – it usually depends how innovative gameplay you want to have, and how good you know your library code. But it IS more assets-oriented and I’m tempted to try this once (especially since nowadays such tools make different platforms versions – also mobile – waaay too accessible).
- See the last round themes? Think about them BEFORE the compo! You don’t have to prepare anything – just think what you could make IF any of these themes got picked.
Okay, now it’s time to rate some games!