Ludum Dare 24 Postmortem

Posted by
August 31st, 2012 10:25 am

Ludum Dare 24 started nearly a week ago, enough time to think it through and
write up the postmortem for “Literally: Earth Shaping”. This was my second LD
and overall I am very happy with how it progressed. The final result was not
much of a game this time, but I am much happier about the process than last time.

What went right

Base engine written before start of compo

I spent the week before this Ludum Dare refamilarise myself with Java and
learning LWJGL. I started the compo with an engine that could draw textured
sprites and play sounds. I could also present menus that produced satisfying
click sounds. This start spring boarded me forward at the start of the compo,it
really gave me a good boost considering the theme.


I discovered Pixen in the aftermath of LD22. It is probably the best dedicated
pixel art tools on OS X. Pixen is quite limited in its ability, as an example
you cant set a colour scheme to use across multiple files. In the end it is a really
useful tool.

What went wrong

Integer Positions

So I built my own engine and it was a really good idea. There was one issue that
cropped up over and over again though. I used integers to store the location of
Entity’s. This made life really difficult when it came to moving things around
the screen. If anything hit (0,0) then it would never move again.

Overall my use of types in java was a constant issue. I used floats all over the
place. In reality floats should have been a cast down from a double in the
OpenGL calls and not used anywhere else. Java is silent about casting between
faults in certain cases, this coupled with the pressure of the compo made life
a night mare.

Update your tools

While Pixen was a brilliant tool for the art I needed it was not at the latest
version until about halfway into the compo. I has a crash just after finishing
the “Play” button and lost about an hour of work. When this happened I decided
to find and install any pixen updates. The updated pixen was a mile better and
I managed to created the play button in about 20 minutes. The lesson here is to
make sure your tools are up to date before the start, you will be able to work
a lot faster.


In the 48 hours I had maybe 3 hours sleep. This is against the popular wisdom,
but I was good on my first LD and I wanted to see how this would go. Starting
straight out was a good idea, but I should have taken advantage of Saturday
night. Sleep is good for you.

Simple Mechanics

48 hours is really not enough time to think thoroughly through and implement
complex game mechanics. I have definitely learnt that lesson now, its just a
shame I had to use this LD to figure it out. I started with good yet complex
ideas, but instead of implementing just enough to get going I did the whole
thing in one fell swoop. If I had only included two or three of these ideas I
could have created much more of an accomplished game.

Sound creation

I really did not know how to use SFXR at the out set and that never really
improved. I also have no real skill in sound creation and so didn’t get around to
creating a sound track. Something to definitely fix for the next LD.

Version Control

I wish I had taken better advantage of version control and other tools. This
would have given me data to create some awesome graphs in the aftermath of
Ludum Dare.

Leave a Reply

You must be logged in to post a comment.

[cache: storing page]