Temple of the Sky People post-mortem

Posted by
September 2nd, 2016 9:52 pm

This entry marks my 5th entry in Ludum Dare, and I think it has been my best so far from an art, game design, and coding perspective.  I wanted to get my thoughts down in my standard Good, Bad, and Ugly format.  I’m not sure about everyone else, but I always go back through and read my old post-mortems before each Ludum Dare to get back into the rapid prototyping mindset, so here we go.

Thog has no time to read your puny post-mordems!

Thog has no time to read your puny post-mortems!

Play Temple of the Sky People here!


I started out with an idea for a story.  Basically a caveman has to search through an ancient temple to find a gem that can cure his son.  The ancient temple was home to super advanced aliens who have this incredible technology, so your caveman would go from swinging a club at the beginning of the game to flying around shooting lasers at the end.  See, it is Ancient Technology, but in reverse!  The ancient tech is actually super advanced!  Aren’t I clever?  Truth be told, I just wanted to make a Metroidvania game and shoehorned the theme in.  Don’t tell anyone though…

Thog, looking at the anti-gravity double jump boots.

Thog, looking at the anti-gravity double jump boots.

caveman runThe Good:

  • The Code – I don’t mean to brag, but my code this time around totally kicked butt.  I went back to HaxeFlixel for this one and for the first time in a game jam I didn’t come across a problem that I haven’t already solved before.  Objects were all self contained and handled their own events so new objects were easy to add.  Everything is structured correctly and I don’t feel the need to refactor anything to work on a post-compo version.  I’m really happy with it.
  • Level Design – I really liked the level design this time around.  I used Tiled to create the maps and I’ve got all the imports and kinks worked out so all the level design work can be done in one place and easily changed, which allowed for rapid prototyping.  The two playable levels in the game were actually created in the last couple hours of the jam.

    All the map editing is done from Tiled and doesn't require any code.

    All the map editing is done from Tiled and doesn’t require any code.

  • Art – I was also extremely happy with the art this time around.  I ended up using the DB32 palette this time around so I didn’t have to worry about colors I chose working together.  Using the limited palette is hard at first, but once you get the hang of it you can create lots of artwork that works together easily.  I had a fairly extensive spritesheet this time around.packed

wallTrapThe Bad:

  • The scope – I actually scoped the game well.  When I wrote down all my items I knew I would need I realized the futility of getting them all in.  So the list of features that I ended up with was actually really close to the pared down brainstorming list.  There just wasn’t enough time for me to implement everything that I wanted to.  I either need to get faster, or not make such complicated games.
  • Time Management – I spent extra time coding correctly, and it screwed up the jam game.  All the nice clean code took much longer than it could have.  I wanted to have certain messages display in certain areas of the map, and instead of showing an invisible sprite when the player got close to it, I instead wrote a messaging system, plumbed it up to Tiled, and created them there.  The 4 or 5 messages you see displayed in the game took over an hour to create.  Great for going forward because nothing is hard coded, but bad for finishing the game in the restricted time of the jam.
  • Music – I still suck at music.  I put together some loops but took them out because the game played better without them.  Because they were bad.  More practice is needed.


The Ugly

  • Final level – The game is obviously unfinished.  I was traveling for work the third day, so I didn’t get nearly the time that I wanted to into the game on that day.  There are large empty spaces in the upper left corner of the map and everything is extremely linear.  The first 2/3rds of the game I really like but the rush is apparent in the last third.
  • Existing finished features got cut – I actually cut some fully developed features.  In the game attacking does nothing.  But actually, in a slightly earlier version of the game you could attack the wall mounted traps and destroy them.  But then I decided that I wanted the traps to also be stepping stones, so I took away the ability to kill them so the player can’t trap themselves at the bottom of a pit with no way out because they killed the step they needed.  I didn’t have time to implement the other enemies, so now attacking is worthless.  The code is in place to allow ducking, but there is no need for it, so I never did the art.  So my planning could have been better because all the jam time I spent coding those features was wasted.  All the code is in place to save and load the game also, but there is no need to save a game that you can play through in less than 5 minutes, so I just didn’t add the button to load to the menu.
  • Starting screen and victory screens – They are just white letters on black.  I put them in early on because I wanted to make sure that they existed but I never got to go back and pretty them up.  Shame really.


Overall, this was a great experience and a great jam.  I’ll be releasing the jam code on GitHub soon when I remember the remote repository commands.  Thanks everyone who has played and commented!

One Response to “Temple of the Sky People post-mortem”

  1. kfischer_okarin says:

    Reminder on the git commands:

    1. Create repository on github
    2. git remote add (remote name) (repository url)
    3. git push -u (remote name) master

    From second push onwards it’s just: git push (remote name)

Leave a Reply

You must be logged in to post a comment.

[cache: storing page]