My next crackpot idea originates in trying to find an optimised way to scan planet surfaces by brute force (you may have seen my botched then corrected calculations quite a few pages ago)
The corrected calculations estimated more than 250,000 co-ordinates to check on the 1st ruins site planet in order to visually scan with eyeball MkI to cover the entire planet - this has now been superceded by another method, which would require a far more manageable 100 sets of co-ordinates to cover the entire planet.
Anyway, the next problem to solve was how to plot 100 equidistant points to check on a sphere. The solution to this requires the use of a fibbonacci series to create a 3D spiral from the north pole, snaking around the planet all the way to the south pole. It does involve a bit of RNG and is an "approximation", however, the nature of 3D geometery and spheres is such that it is the only practical way to do it.
My crazy crackpot idea is that I need to find the "magic" number of points to generate for any given planet (there are many possible variables), and that one of the points generated by the fibbonacci alogrithm will match the coordinates of the already found sites.
I will be writing the programs to work all this out over the next few days, and running the numbers against known data. My inspiration is that ProcGen and RNG are involved in this. RNG based on fibbonacci series is something DB has talked about in the past - it's a guess, a leap of faith, a gut feeling, nothing more than that - but I feel worth my time looking into for a bit....
The corrected calculations estimated more than 250,000 co-ordinates to check on the 1st ruins site planet in order to visually scan with eyeball MkI to cover the entire planet - this has now been superceded by another method, which would require a far more manageable 100 sets of co-ordinates to cover the entire planet.
Anyway, the next problem to solve was how to plot 100 equidistant points to check on a sphere. The solution to this requires the use of a fibbonacci series to create a 3D spiral from the north pole, snaking around the planet all the way to the south pole. It does involve a bit of RNG and is an "approximation", however, the nature of 3D geometery and spheres is such that it is the only practical way to do it.
My crazy crackpot idea is that I need to find the "magic" number of points to generate for any given planet (there are many possible variables), and that one of the points generated by the fibbonacci alogrithm will match the coordinates of the already found sites.
I will be writing the programs to work all this out over the next few days, and running the numbers against known data. My inspiration is that ProcGen and RNG are involved in this. RNG based on fibbonacci series is something DB has talked about in the past - it's a guess, a leap of faith, a gut feeling, nothing more than that - but I feel worth my time looking into for a bit....
Last edited: