Reddit
Postmortem
First things first: I finished, and that is great: http://www.ludumdare.com/compo/ludum-dare-24/?action=preview&uid=3996.
I aimed for some tower defency thing and decided to learn some svg and svg animations this time (it is allways good to learn something). After the first day, I was afraid I would not be able to finish, because I got stuck with how svg animations are dealt with by different browsers, especially when using the same animation multiple times. I decided to drop safari support, sound, and fancy graphics. Keeping the prototype graphics seems to have been a good decission, first commenters seem to enjoy it. Safari support could be added later. Sound will come, when I do post-compo improvements. So I’m sattisfied overall.
I personaly like it how the creeps get better over time, but you can not quite put your finger on it, when it happens. To me, this feels much like how darwin explained his theory of evolution. Originally, I wanted to evolve the pathfinding, too, by tweeking the parameters of the algorithm, but I dropped it because time got thin and I figured most people would build one-way mazes, so it would not make any difference. Maybe I try it when doing post-compo things.
What went right
- I got the idea quite fast, only 1 or 2 hours of thinking and some sleep
- The simple graphics are minimalistic and not loosy.
- The simple mechanics provide a quick learning curve, which is perfect for LD.
- I actually implemented creeps, that “learn” how to break through your maze.
- I made it, allthough I got a normal amount of sleep.
What went wrong:
- Performance of svg rendering and animation varies largely between different browsers.
- SVG animations, that are triggered by javascript, are evaluated differently across browsers. This made me loose a lot of time and ended up in detecting, which browser is used. Details can be found below.
- No Sound. This annoys me especially since I had sound in all my other LD games so far.
- No decent graphics. Well turned out not so bad
- No balancing. Seriously, I have got no idea if the game allways gets to end, or if you can reach a point where your towers get better and better and you can play endless. I hope that is not the case, as towers get more expensive on a cubic curve while the bounty increases linearly.
- No win condition, No highscore (I should have set up a server side scripting language and a database server before LD)
- I got multiple little fixes I delayed and forgot, I should keep a list next time
- One player said, he was not able to start, I hate it when this happens
Details to the SVG animations:
- FF and Chrome do not actually change attributes but apply a transform, but each animation is reset before using it again, so I still need to change the attributes manually afterwards and can not read them in between.
- Safari is not able to reuse animations at all, Once ended, it can not restart. So I needed to clone the element each time and start the new one. But it actually changes the attributes (Well, I think, I didn’t take a deep look into it…)
- FF supports the onend event, that fires after the animation finished. Chrome and Safari do not, I used timeouts as a workaround.
- The method “beginElement” starts an svg animation from javascript. In FF and Chrome, ALL animation elements support it, in Safari only the animate element does, animateMotion for example does not, so I could not start it from javascript and could not animate the bullets in Safari (It just shows a line to the target instead).
Tags: postmortem