Here's a scan that discusses a missing abstraction in my model so far, to carry on a "thread" on a map location (as compared to a tileset location). I figured out I was missing it when I tried to implement crumbling floors for the "nut's'blots" intermediate project, in june 2010. It remained a draft sketch all that long.
An amusing fact: despite this is still missing from the *model*, there is actually creation of a GameObject derivative (BlockArea) so that you can proceed with evaluation of the "block:on hit" expression ... so all it would require would be to capture this derivative for longer than just a collision. I just realised that while investigating why the new collision system no longer works with interactive blocks...
Un vieux scan de Juin 2010 avec, en Français, le détail des "BlockAction" qui manquent encore à mon modèle de jeu. D'une manière assez amusante, la manière dont les interactions sont gérées vont déjà dans ce sens même s'il n'y a rien au niveau du "gobscript" qui permette d'en exploiter plus largement les possibilités.
A less amusing fact: this was all coded before we had dynamic gob lists, and it's now trying to de-register a Gob from the game engine several times every frame (especially when the interactive block is not de-activated by the collision). That could explain some performance issues observed in early AppleAssault prototypes.
No comments:
Post a Comment