Link to my LD-Entry Firefly
I don’t want to present a list of good and bad things, so I’ll take a bit more of a narrative approach to my post mortem (although english is not my native language)
The Idea – and how bad at drawing I am
This is my first Ludum Dare entry. I didn’t really know what to expect from myself, and the first few hours after theme anouncement, I was constantly switching between a “lets try just to have fun” and a “Lets see what is the best I can do in 48h” mood. I ended up somewhere in between and I felt very good doing things as I did them.
I had a very precise idea for an evolution based theme in my head but I took some time to make up some others to see if I could find a better one. The only one that I really came up with was “The Darwin Award Game – Kill yourself as dumb and fast as you can” which I gave up on once I told my girlfriend about it and she only shook her head, saying nothing. So I decided to go for the “plants and bee’s”-theme.
I wasn’t really sure what the final mechanic should be like. I only knew that one fundamental process should be part of the game-play: Bee’s pollinating plants and mixing genes following rules similar to mendels-rules.
So the first thing I did was drawing a sketch of what a game-screen could look like. Here you see a comparisan graphics of what the sketch, and what the game that I finally submitted look like: (I am not good in drawing obviously, but for the amusement I’ll add the sketch anyways – and hey! It was 5 AM!).

The Programming
I had decided the day before theme announcement, that I would start with very sketchy placeholder graphics and spend the first 24hours coding. I would then start to create nicer graphics and add sounds to finally do some polishing. I stuck to that plan and dived right into an empty eclipse project that I had prepared with the libraries that I chose (Libgdx and Artemis Entity-System-Framework). I planned to submit an Android and a java version of my game. That was important to plan upfront because I had to map all interactions to touch-input (at least in my head) when designing controls. Artemis proofed once again to be a perfect match for rapid development like it was required for LD. But I really feel very comfortable with libgdx and artemis anyway. (I ended up messing with the Entity-System pattern a bit when I had no time left and produced some not so good looking code :))
I think it helps a lot not to spend all the time with a new api or tool you haven’t tried before. Why do so many entries choose tools or libraries they never used before for a LD entry? It doesn’t sound like something fun for me (unless you need to give yourself a handicap because any bunch of numbers turn into an incredibly looking, sounding and playing game if you only look at them).
Sunday morning, I finished the last systems I felt mandatory to be present to… yes … to what? I realized in that moment that somehow I was still not really clear on what the gameplay should be like exactly. What I had was:
- Plants, growing blossoms (not really because the placeholder graphics didn’t look like it, but I knew what it could be like).
- Three slots where you could put a plant.
- A weather-system which would randomly spawn snow, rain or heat.
- Plants could die from weather that they were not resistant against.
- Bees were flying from plant to plant eating their pollen and pollinating other plants
If you played my game, this might sound familiar to you, because this sounds very similar to the games tutorial – for a reason.
The Game-Mechanics – or how I would first do something else
Not many things changed concerning the game play. Things that I introduced sunday morning to have at least something to work with was:
- The pollination process got a bit more complex (Mutation was added for example).
- Generic-Seeds that would give you some start-stats and are required to grow plants with new random stats were introduced.
- Blossoms were made harvestable after being pollinated so you could plant them and grow a new plant with the blossoms stats (which are independent of the plants base stats after being pollinated).
- Bees were renamed to fireflies (because bees suck).
- Bees could have stats attached when they entered the game that were random (to bring a bit more diversity into the whole setup).
- The player was able to smash fireflies if he didn’t want them to pollinate a plant with stats he didn’t like
That’s basically it. I didn’t think about what the gameplay would be like anymore because I couldn’t spontainously come up with something and just started to hope that I would stumble about a brilliant idea in the process of creating the assets. So I started creating graphics (I did them all in inkscape).
This is what the game looked like sunday early morning:

I started to work on the plants and played with different background graphics. It was fun to do, but I always had in mind that I still had not the perfect game-play that I wanted.
Another step in the process:

So you have seen in the beginning of my post what the game looked like in the end, so I won’t insert more screenshots. (Pro Tip: To find out what it really looks and sounds like, it’s helpful to play and rate the game ;-))
I finished most of the graphics around 5PM on Sunday (around 10 hours before the deadline) and started to search for the cables that I needed to record from my e-piano to my computer. That took a while (I drank far to many energie-drinks by that time – thats really a don’t. I could hardly concentrate and giggled like crazy for a while…). Once I had calmed down again I recorded the music for the game (that was really a quick thing, I knew I didn’t have much time to work on it because there were several other things to be done). An hour later I had the music cut, and written a system that would load and play them in the game. I created some more sounds for selections and mutation on plants but ran out of time quickly so I couldn’t record sound effects for the weather or the fireflies for example (Would have liked to do that).
The Wake-Up-Call
Once these things were ready there was not much time left. Maybe 3 hours. I was really pleased with what the game looked like, given my drawing talent and the short time. But I was (and am) very unsatisfied with the gameplay. I mean its all fine, once you read the tutorial or found out what possible interactions are, it’s really playable. There is goal to achieve and there is actually quite some complexity involved in it (The goal is to cultivate a plant that survives 5 minutes) But when I was play-testing, I could hardly achieve that goal. There were two reasons for it:
- The feedback loop was hardly present. As mdkess wrote in my games comments later: It was hard to figure out how good you were doing.
- The game was actually to hard. The pollination process was rarely beneficial and you had only a low chance of a mutation to be beneficial either.
But the deadline got close and I knew I had not enough time to change the game to improve point 1. So all I could do was to insert an easy and a hard game mode. The hard game mode has all difficulties left as I describe them here. The easy mode really is much easier, so if you want a challenge, try the hard mode (though I suggest to start with easy mode to get to know the game). The other thing that I put into the game within the last hour before the end of the compo, was a tutorial. I really don’t like what it became like. Obviously I couldn’t integrate an interactive tutorial (which I usually like to do). So its just 10 screens filled with text and some spare graphics to explain everything. There is now a tutorial outside of the game that is a bit cleaner and easier to understand. I suggest to read that one 😉
So I guess now that I wrote so much about it I actually want to do a list :D:
What was gooooood
- Schedule was nearly perfect. Although some things are not as polished as I like them to be, I feel like I really spent my time on the right parts to have a full game done by the end of 48 hours.
- Programming was great. I really did some not so pretty things shortly before submission deadline, but that was not avoidable to get done.
- Idea to game conversion worked really good. Comparing my sketch with the result its actually not that far away from each other.
- Graphics and sounds are much better than I thought I could do.
- Porting the game to android was a matter of 60 minutes. Planning for an android release really paid off.
- Really not many bugs throughout development. If there were some, I found them fast and it didn’t take long to fix them.
What was baaaaaad:
- This whole game-play thing really sucked. Now, a few days later I really think the game-play itself is great, but it’s the lack of communication to the player that prevents it from shining.
- Too much caffeine. I know, there isn’t sth. like too much caffeine – but – yes there is.
- One of the libgdx extensions I used (freetype) has a bug which led to most of the first visitors of my game not being able to play it. Took a while to figure that out. In the end it wasn’t my fault, but I hated that people who wanted to try my game were unable to do so. (It should all be fixed now).
Can’t think of anything more right now. Maybe I’ll remember sth. later and add it.
So much for my first Ludum Dare. Thanks for reading this, hope you found at least parts of it interesting.
Now go and play it if you haven’t 
Best,
Tim
Link to my LD-Entry Firefly
this is like reading harry potter!