Talk:Flawed collision detection
Should we be treating this like it's commonplace, or not?
As I slowly grind my way through the COMPET-N archive, I've been noting really blatant instances of this bug in the corresponding walkthrough article. It now occurs to me that this might contradict our notability policy, but to decide that requires more technical expertise than I have.
The policy advises not writing about "how a particular bug affects a particular map, when the bug and the map already have articles." In other words, no individual sub-articles for each map containing both an invulnerability and a sky. The difference is that any idiot can identify such maps, whereas this bug apparently does not screw up every map that has a BLOCKMAP lump, so more subtle reasoning is needed.
I can't help thinking that, in principle, the bug could occur almost anywhere, because no one actually takes it into account while designing a level, and therefore the geometry of walls and passages is uncorrelated with the blockmap one way or the other (except possibly by ludicrous coincidence). Might there be, on the other hand, statistical methods for deciding that players and monsters are more/less likely to approach some intersection or room from certain angles? That seems to be micro-analyzing people's movements beyond any reasonable expectation of utility. Yet there do seem to be places where competent players have encountered the bug repeatedly (IIRC, besides cph's marvelous narrative, the first item in my E1M7 list has recurred periodically).
I'm going to shake my head in calm discontent and keep noting blatant instances in walkthrough articles, but I thought I'd post here in case I'm missing something trivial, or people with programming knowledge have thought about this already for some reason, or it's 30 years from now and this kind of esoterica is the only thing we're still actively debating. :D Ryan W 10:51, 16 September 2007 (UTC)