No actual update, just some assorted discussion. Not that anyone is seriously reading this blog^W random crap dumping zone, much less for a 3 year old free time project.
Honestly some of the most terrifying maps
Hahaha yeah. Definitely harder to get a good circuit or line going
When it’s all CHAOS
A: I think a bit more cover would be useful
A: But it’s been so long it’s hard to tell lol
H: Maybe future ones will tone that down and produce more normal stuff.
H: E.g. big central buildings and streets kinda things rather than just messes
A: Eh. Messes have merits. I think map design is just hard
And the things you want to optimise for are so hard to simulate and score
I wonder if there could be a line of sight test when attempting to place a room?
LoS is harder but. Yeah,? Whatcha thinking
Like max line of sight distance to make sure there are hiding places?
…
So in my thing most rooms in my thing have roughly a transparency, either they’re empty or they’re solid or partially solid. So you could calculate that for every position - every 15 degrees around in a circle.
You could score it. Yeah
But integrating at runtime and having it make sense would be harder
Easier to post evaluate
and a counter point
H: The one with the tubes would pass with flying colors there
H: But it wasn’t fun because you couldn’t see anyone until you were on top of of them
A: True. That one was crazy
H: So maybe needs a min and max distance
A: Tubes aren’t bad but maybe a max length
on placing pieces, should definitely have a ‘preview’ step in the pipeline, each room type gets a single voxel, and run our tests on that or throw it out. constraints maybe useful here?
H: And was always tough for me that the big sexy pieces were hard to fit in there without manually placing them
H: If i added a 2 unit wide tunnel for instance, I’d make it 2 units by 2 units, and maybe a plaza on the end. But then it never gets placed because it’s huge and unless you start generating / growing from the edges you can place fun stuff easily.
H: But that’s def a problem that has other solutions. Like generating more … Like map previews and tossing out boring ones before generating the real ones?
If we have another path method
A: I wonder if you could randomly place 2-5 big sexy pieces and then generate paths around them to ensure patching then randomly generate rooms in those paths
A: And fill in around that
H: Yeah, so that’s another strat but I don’t actually have a reachability checker. Ieverything was reachable just by nature of “pick an end and extend.”
H: We’d have to add a validator step to ensure everything can be reached. Which would be useful to have
on CTFs:
H: And never buult good CTF maps and i want more of those. There’s fun stuff you could do… If there’s multiple paths between each flag room. But that requires v diff algos to ensure multiple psths
H: More of a “design a path and then find rooms which fit” rather than the current “just make random paths”
A: CTF is a whole nother ball game from DM
A: Yeah it’s like make 4 paths between two poits
H: DM is more fun if everyone is forced into a small number of arenas too. Which isn’t a thing currently
H: Also currently it mandates that things are walkable, we ignore parkouring. But CTF could be fun if you’re forced to take a diff path back or so