@foggygoofball in that case i wait until the next build for the next publicity wave …

@steve I know you’ve gone camping, so I’ll do my best to work it out for myself, but I’ve got a question about notes you made a dozen or so years ago.

<Steve>Creatures need to be able to learn how to navigate around their world using landmarks (something even some insects can do). Their main visual system has to be limited in range or else they’d have to do ray intersection calculations with hundreds of objects many times per second, so how do they see landmarks like distant mountains, lakes, buildings and so on when they’re short-sighted? That’s partly why I decided to separate this problem from the others. The world is filled with objects that are marked as landscape features of one of a few basic types, and the creatures directly sense the absolute bearing of the nearest landmark of each type. If I were being fancy I wouldn’t tell them the absolute bearing – I’d let them see where each landmark is in relation to their eye and their brains would have to combine this with knowledge of where their eyes, head and body were pointing in order to work out where in compass coordinates the landmark was. But since I’m cheating anyway, and since their brains already have to go to considerable lengths to do this for their normal vision I decided it probably wasn’t worth adding such unnecessary duplication, and so I just tell them where the landmarks are as compass bearings.</Steve>

Does this still hold true in our current model? Landmarks are actually like telepathic lighthouses broadcasting their coordinates to all of the brains in the world?

If I understand it, it’s less like a cheat and more like how migratory creatures have been shown to sense magnetic field maps and such.

Maybe that technique is a cheat, nature cheats all the time, right?

@foggygoofball my ducks defendly cheat that way, gave some drakes away, they walked 3 times om the road home until the hearding dog captured them again. Always in the right direction, never in the wrong direction…. always with 3-10 cars slow behind them…

So upon further study I believe I’ve discerned that the landmarks aren’t given as compass bearings anymore, they are instead mapped out as a (hopefully) unique combination of values (still haven’t quite gotten to figuring out how they’re assigned yet, maybe I’m missing it and they actually are compass bearings in retrospect) assigned to every grid point of x and Y across the entire map. 

I understand the 3×3 grid of local space that the eyes and ears utilise to pinpoint stimulae, but now I’m left with two new questions.

Since the visual range of the creature determines the actual size of individual zones and therefore the physical dimensions of the world grid; A) does shrinking the visual range and thereby increasing the number of grid nodes in the world cost significantly more CPU time, or even the other way around?  By widening the VR we encompass more potential objects for the daemons to assign salience to, presumably this would increase cpu usage as well.  There must be a sweet spot for performance methinks.

And B) does each creature create and maintain their own instance of the world grid and all it’s landmarks beyond the 3×3 local grid space, or is the same map available to all creatures of the same VR as an intrinsic property of the world?  Calculated only once and then passed on to the individual brains as it were.