Galcon 3: the third -- join the party
 
 

Classic :: Post Mortem

Home | Download | Buy | Forum | Net Ranks | Support

In the spirit of the Ludum Dare, here's my compo post mortem.  Like hamumu, in my timezone, the compo runs for three days -- one late evening, one day, and one day minus the late evening.

Day 0: Preparations (before the compo began)

I competed in the Ludum Dare warmup compo the week before, creating a rather boring "vector" game, as that was the theme.  My game wasn't even really a game.  My usual language of choice is python, but for this compo I wanted to step up the graphics a notch or two by using something other than python -- C, of course.  Unfortunately, C uses { } and other non indented things, so during the warm up I spent most of my time building pyplus an "indented C/C++" to C/C++ converter.  Makes coding in non-python a bit more of a joy for me.

I was pushing for a number of themes Adventure Games, which won the first round of voting, but horribly lost the second round :( ... I thought it would be a good challenging theme, even if it would likely yield very few "good" games.  I didn't really want swarms because only a month ago I made Zanthor with illume and some other friends.  The game contained swarms of peasants which you destroyed with cannon balls from your rickety flying machine.  I was a bit swarmed out.  (As is that came as an advantage, I knew how swarm algorithms worked in advance, so I didn't have to spend hours researching and figuring out the details of them, since it was fresh on my mind.)

I did some work to setup my win32 system so that I could release a py2exe version of my game as well.  Got mingw and SDL and all that installed more or less so I could do my C graphical extensions to pygame.

Day 1:  Nothing

Day 1 was spent in my usual panic mode.  "Oh no we have a theme."  So I frenzied around my house thinking of swarm games and asking a few other people what they think and get all kinds of dumb ideas and blah blah blah.  I go to bed, having accomplished and thought of nothing.

Day 2:  Get to work!

I got up on Day 2, quite refreshed from a nights sleep where I didn't think of any good theme.  But I decided that as was my original plan, I wanted to make a game with good graphics.  I really got a lot of that inspiration from one of bluescrn's post mortems, where made some comment like, "I just come up with some cool tech then add the game on at the end."  Since he usually faired pretty well, and I felt like doing better in the graphics category (not usually been my best category, since I usually don't spent loads of time on the graphics).

I decided to make a space conquest game, maybe a bit like Galactic Conquest a DOS based game made in 1987.  It was pretty fun.  I figured I could do a re-make with really cool planet graphics and maybe swarms of space ships or something.  Thus I decided to spend the day making whatever look good.  Here's a timeline (from looking at my svn log, I can recall what I did when...)

9:45 am - inital import
11:14 am - plasma algorithm
11:54 am - basic planet generation
1:29 pm - planets with atmosphere
2:00 pm - better looking planets
4:23 pm - particles and ships
4:51 pm - background generation

I was quite pleased with the basic game graphic elements I had so far, so I moved onto getting a basic game working.  And then I went to bed.

12:09 am - basic game working

Day 3 - Polishing it up

9:30am - 12:00pm - I spent the morning doing lots of miscellaneous fixes.  Graphical touch ups, basic gameplay tweaks, etc.
2:00pm - computer AI
4:59pm - menus and prompts, and all those niceties
5:32pm - music (composed in modplug)
5:58pm - sfx (recorded with a fostex MR-8, touched up with audacity)
7:28pm - tutorial
uploaded game to LD site

Day 4 - Post Compo

Fortunately, the LD rules allowed me to get my win32 port working the next day.  I found that my C extensions did not want to cooperate nicely with py2exe for some reason.  It took me quite some time to wrestle them into working, but eventually I did have a win32 release.

Good and Bad

Most stuff went pretty good on this one.  One thing that went particularily well was having my friend Tim do play testing during Day 3 for a while, he's quite good at knowing when a game feels right -- so I just keep coding and he keeps writing down little tweaks and comments and eventually the gameplay balance gets nice.  This is probably one of the more important things to do to get a good game, and it doesn't take much time :)

The Bad part was the win32 port.  I should have prepared for that better.  In the future I'll make sure everything works in advance before I go and assume it does :)

Not much else to say.  I was quite suprised to see how well using bluescrn's LD strategy worked out for me.  It was quite fun too, since usually I just get some rough looking graphics together and whatever, try to get a game out of it.  This time I got to have really fun looking game assets to work with once they were created.