Star Tribe PM

Posted by (twitter: @@grimgary)
April 21st, 2016 6:24 pm

Post Mortem on my game Star Tribe.

About three weeks fore the LD weekend, a mental reminder went off in my head.  Check Ludum Dare dates.  I quickly asked my boss for those vacation days.  It had been a while since my first LD some few years ago (10 Seconds Theme), and I wanted to try it again.  I had tried about a year ago without taking time off, but this didn’t work out so well. With work, I just didn’t have the drive or energy to do a 48 hour game no matter what the size or scope.  But this time I was prepared.

So the weeks before the compo I imagined the type of game I wanted to build.  I had been looking at the old Star Control games and figured a space shoot em up like that would be a fairly simple reach, and could be adjusted in theme when it was announced.  So I had a plan in place.  One that I ended up not sticking too, and much to my detriment.

After the theme was announced, I hit the ground running and started with some sprite concepts.   I could absolutely do a space shooter like Star Control battles.   Just needed to make the ships transform.  And for some reason native American themes had been roaming around in my head.  My original plan was starting to mutate already, setting me off track.


Eventually I ended up with four designs for a single ship in the next few hours after the theme commencement.


And those are what I stuck with.  I could have easily stuck with a single ship, with a rock/paper/scissors format of battle.  But I didn’t.  I then began to imagine a small squad of assistant ships to help each side.  Then it shifted to having an invader fleet against a desperate native population woefully unequipped to deal with a large unending wave of invaders.  So the enemy became carriers with huge compliments of light fighters and the player would command scattered small squadrons called war bands.

By Saturday I was into the code.  But I hadn’t finished all the art I wanted to do on Friday.  I figured I would relax and contemplate what my next moves would be.  I spent fourteen hours in total on Saturday constructing a simulated off screen battle system…which eventually got erased by Sunday.  Instead I opted to have all fighters be presently spawned and have the AI be calculated for hundreds of various fighters on each side.  Which, on my system was working fine.  By the end of Saturday night, I had the enemy and friendly forces able to engage each other at the very least, but so many elements were missing.

Here is where my problem started to come in.  I wanted to complete the game for the 48 hour compo.  But on Sunday I was running out of time.  I would have to shift over to the 72 hour jam.  No worries.  Crossing the finish line with a viable game seemed to be the goal here.  BUT.  For every new feature or set of features I had to test if they work.  In the face of time constraints, having a game that takes a long time to play for testing will kill you.  A play session could last for 30 minutes or so (well, it’s endless technically – but to do a full round through the features takes a while).  I had yet to finish some minor graphics, make my audio queue for sounds, and did not have key mechanics in (death for instance).

Sunday came and went with another 14 hours or so of work.  It was looking good, but I missed the 48 hour deadline for sure.  I was almost depressed.  But I wasn’t out.  I could still complete for the jam, which also allowed me to add in some cool music from Kevin McLeod of Incompetech.   Come 9pm, I missed having audio in the game.  Barely got lose conditions in the game, and didn’t have any win conditions.  However, it wasn’t a bad run overall.  Just a bit too large of scope.   It became a great concept for a larger game.  One that will likely never happen.  But I will add to game as time goes on, just not enough grand features to take it too much out of the way of being a space tower defense game of a sort.

I will likely stop using BlitzMax after this point.  The engine is fine.  It is still, after all these years still quite capable of being fast. But only when you optimize your raw code for your own engine.  Which, who has much time for organizing and assessing the bulk of code for optimization during a time limited development jam? Not too many people.  So I suppose it may crawl on some people’s systems due to the way it is written, not because of the tool.  However, Blizmax at this point has limitations that is drawing me away from it anyway (Deployment targets being one of them, and inability to easily create/manipulate rendering surfaces without going through some serious long winded GL hoops).

I should have just stuck to my original plan.  I’d likely have made the 48hour compo time limit easily with two 14 hour work sessions and made a pretty decent game in the process.


Leave a Reply

You must be logged in to post a comment.

[cache: storing page]