karl3＠writeme.com wrote:
karl3＠writeme.com wrote:
karl3＠writeme.com wrote: karl3＠writeme.com wrote: karl3＠writeme.com wrote: karl3＠writeme.com wrote: <! ^(@/ machine learning marketer: "i'm not sure if the term "slave boss" is really ideal for you ..." slave boss: "ah, yes, the social programs i've paid you to instigate are working already! yes. nobody says "slave boss". when they say that i kill them." machine learning marketer looks at slave boss o_@! ? karl3＠writeme.com wrote: the quote levels don't work quite right in this hyperkitty 1.3.12 mailman interface. i have an even better idea than an n-dimensional maze solving algorithm in 4 or so lines! [ ... it's an old idea -- let's figure out a way to have _partial dimensions_! fuzzy dimensionality! vagueness, uncertainty, tangled wrong-like parts! what would a maze-solving algorithm look like in a sloppy mix between 2+1/4 and 2+1/3 dimensions? maybe a rough bell curve distribution centered aroudn 2+7/24ths dimensions? what would the maze representation itself look like? there are so many different possible approaches! -- attempting to send the above email gave the web2 interface error "undefined". i'm waiting a bit and checking the archives to see if it went through quickly. 719p et. 720p not seeing it. sending again feeling out how this lands (yes i suppose we could have only some avenues that reach into higher dimensionality, and even make the dimensionalitiness arbitrarily and infinitely high but use it in a distribution that provides for finite time (somewhat interesting!) but) found a debateish over here, this doesn't seem quite right to some. if there are infinite dimensions, how can it possibly be in finite time? but of course i meant finitely-dimensioned hyperplanes in infinite nominal dimensions, which isn't nearly as fancy but sounds so exotic and esoteric that it's hard to differentiate you like the idea. you want to try it out. hrm. it's not as fancy as it sounds. let's see if i can summarize it ... [[[[[ basically you define a distribution with an infinite tail that decides how high-dimensional you will be at every turn in the maze while cutting the paths. i don't know the correct math terminology, but a distribution such as `f(x) = y = rand(); y < 0.5 ? y : 0.5 + f(x) * 2` is potentially infinite when written without mistakes (i likely made a mistake writing it), as well as practically always evaluates in finite time; i might spend some time considering the space to adjust the curve and speed you'd use infinitely-large-but-usually-small distributions to decide what dimensions to engage in each step of cutting the maze you'd likely place bounds on it so that the maze closed; this is kind of like modeling a fractal or limit system; but you'd get used to it while doing it and figure some things out for example 3-dimensional portions jutting out of a 2-dimensional maze will likely want to return to the 2-dimensional plane to keep the maze interesting, so the choices of paths could be weighted based on that you could simplify it to an expansion on the drunkard's walk such that the drunkard can invent new dimensions if he feels so inclined hmm that didn't go quite as planned. maybe wasn't the best exploration. but it wasn't that bad. i think maybe i'd like of course to expand the idea of a maze into the space in which it is represented and held, kind of. how could the dimensionality and path-connectivity-meaning of the maze, itself be maze-like? now i'm having a little trouble remembering my preferred idea (as well as noting the other idea of interest got somewhat injured). let's go back to (a) mazes that (b) are more confusing than normal by applying the concept of maze into a further space of the concept of making a maze or what a maze is ok i'm kind of thinking, for each step in the maze there is a "maze form" that can change. this is like a new dimension except it can be something other than a dimension: for example, there could be a network of portals, or you could change into different people, or the maze could be re-interpreted based on a pizza flavor. maybe you have to solve a riddle and which solution you pick decides which door opens. i'd like it to be more than just "links", like maybe the whole idea of the maze could change sometimes so that it's not apparant any more that it is a maz-- [but of course in the new scenario you could (or might have to!) return to it being a maze (so as to reach the exi--
and now we get to work on the multitimeline simulator again :D i started looking for it on github but am now experiencing a physical world situation