It’s quite possible Lime had a lot of hand-holding features, but as of 3.5b most seem deprecated or undocumented. I think you have a great opportunity here because the framework has petrified so much that you can make changes at will; personally I don’t think Legacy is a big deal. Leave 3.5b access to anyone who needed it, and then make 4 the fresh new start.
The step too far I think Lime took was in trying to wrap too much of the Corona pipeline in itself. Spawning objects according to Tiled objects, for example, quickly fell out of favor because users have other methods (like imageSheets) to consider. Lime needs to be flexible; it needs to bridge the tile background/foreground world with the rest of the Corona engine, but ideally it shouldn’t attempt to replace it - if it replaces too much then every time Corona Labs makes a change Lime will be broken and lagging a step behind.
(I’d point out specific features but as the website is now gone, I really have no idea what to point at…)
Lime’s roots should be:
(1) A fast, working tile engine that supports very large environments. I think it’s reasonable to assume the iPhone 4 becomes the bottom platform next year, so that’s a reasonable target for performance/memory constraints. If you can’t run the Dragon Quest 2 world map (256x256) or a Super Mario Bros. level (187+x13) there’s a problem.
(2) A helper for common tasks related to the tile engine, like camera control, snap-to-sprite, layering, possibly parallax, etc.
(3) An interface between Tiled and Corona. Tiled was one of the best decisions of the software because it’s multiplatform, open-source, and it means you don’t have to get your hands dirty writing/maintaining tile placement software.
Other Wishlist Features
(1) World Wrap : Huge for RPGs; just need to connect the map edges.
(2) Trick Danny into supporting Nearest Neighbor OpenGL Rendering : For 8bit fans… [import]uid: 41884 topic_id: 33825 reply_id: 135120[/import]