LÖVE

I recently played Robot Wants Kitty (I recommend you to play it too ), a "metroidvania" platformer, which means that the game consists of one huge level, where you can get powerups like double-jump, or shooting lasers, to be able to get to other parts of the level. This inspired me a lot to make something like this myself, using bump. But, I have a question about this - should I make the world contain the whole level, or separate the world into smaller parts that will be loaded every time the player gets to the edge of the screen (this will not be considered different levels since the loading happens simultaneously without any timer)? I don't know if bump can contain so much data at once (including moving platforms and moving enemies).

Doing one big world shouldn't be a problem. Bump uses "cells" of a certain size to check for collisions between objects that exist in the same cells. So if you have one million objects and they're all moving, but none of them every end up in the same collision cells, then their will be no collision checks going on.

Sir_Silver wrote: ↑Wed Apr 26, 2017 3:08 pm Bump uses "cells" of a certain size to check for collisions between objects that exist in the same cells

Ditto on that, I would be very surprised if bump was a performance bottleneck for doing it all in one level. It uses a very cool cell-based system to achieve high performance.

If you're using the STI library for the map rendering, Karai did some performance optimizations back in February (https://github.com/karai17/Simple-Tiled ... a05ac7fc9a), but I don't know for sure if it will be happy rendering a lot of layers of a truly massive map... it's using quads and sprite batches, so it's very possible that it'll do just fine, but I don't know for sure.

I suspect that the usual programmer advice applies, to not do prematurely optimize. IOW, wait until you know that it's a problem before investing the time & introducing the complexity of splitting things up into separate maps.

Thank you for your replies!