It seems to me pretty unlikely that most borders regularly pass exactly on the near infinitesimal point represented at the corner of two pixels.
That implies then that the rule for determining the colour of a pixel is not: in this square is there any border?
So what is the rule? If it percentage based then it would run into problems when the line does match up with the border, for example if you said that 10% of the land has to be in another country you could end up with two areas with no countries touching.
Similarly, if we imagine there are two basic pixels for a map: land and water (at low tide? at high tide? another but not relevant issue). If you look at a tile what percentage of the surface has to be land for it to be land or water if vice versa? When do you use the coast tile? When there is less than 90% land and less than 90% water? Again this could lead to a region of no border tiles, is that fine? After all if we use these rules anyone who looks at the pictures and knows the rules will instantly know that the border of a place with no pixels must be on the edge of the pixels.
Right now, if you look at a pixel? What can you say is true? Can you say there's no border, well apparently not because the borders all use diagonal neighbours. Can you say there is no water? Well no, because coasts follow the same rule and I have no idea if coasts favour land, or favour water, or randomly favour one side depending on the whims of the map maker.
If you had an AI who could examine every grid square on earth what rules would it follow? Presumably it is necessary for understanding each others maps that these rules are established and clear to everyone, otherwise new map makers would have maps misaligned or in disagreement with other maps. But I've never heard of these rules if they exist, how do people know when making maps what colour the pixel should be? Especially when making new maps, but also when mapping modern external and internal borders?
Edit:
An example of where this matters is the portrayal of microstates. Some people place a coloured tile surrounded by four border tiles, but that makes no sense. Most microstates should simply be a single border tile, or two if they lie across the pixel line, or four (in a square) if they're on the corner. They should have no internal colour.
Maps that would give them a colour would need more than 200 (maybe 700) million pixels [20000 by 10000] as there are 200 million square miles on earth and the Vatican is 0.29 square miles,
Edit 2: I tried to line up the Netherlands border from 2019 QBAM to see if there was any coherent pattern:
I spent ages adjusting the scale and moving it and this was about the best I could get. There seems to me no coherence here. You don't expect it to be pixel perfect but that there is a gap at all instead of an enclave seems really difficult to believe from the relative position of everything else. If you thought well at least I know from this map that all the territory within this area is netherlands you'd be wrong on that, but it seems to me the whole point of a border tile is for when both countries have territory in the same pixel. If something is coloured orange for the Netherlands shouldn't it be completely Dutch?
Edit 3:
This is also very important when discussing sea level rise and fall, or changes to rivers, reservoirs, lakes (Aral, Caspian, etc). At what point will the pixels in the Pacific be changed from land to water? When the islands are fully submerged?
Edit 4: An interesting way of thinking about this problem. If you had to map the earth with 1 pixel, what colour would it be? What about 2 pixels?
That implies then that the rule for determining the colour of a pixel is not: in this square is there any border?
So what is the rule? If it percentage based then it would run into problems when the line does match up with the border, for example if you said that 10% of the land has to be in another country you could end up with two areas with no countries touching.
Similarly, if we imagine there are two basic pixels for a map: land and water (at low tide? at high tide? another but not relevant issue). If you look at a tile what percentage of the surface has to be land for it to be land or water if vice versa? When do you use the coast tile? When there is less than 90% land and less than 90% water? Again this could lead to a region of no border tiles, is that fine? After all if we use these rules anyone who looks at the pictures and knows the rules will instantly know that the border of a place with no pixels must be on the edge of the pixels.
Right now, if you look at a pixel? What can you say is true? Can you say there's no border, well apparently not because the borders all use diagonal neighbours. Can you say there is no water? Well no, because coasts follow the same rule and I have no idea if coasts favour land, or favour water, or randomly favour one side depending on the whims of the map maker.
If you had an AI who could examine every grid square on earth what rules would it follow? Presumably it is necessary for understanding each others maps that these rules are established and clear to everyone, otherwise new map makers would have maps misaligned or in disagreement with other maps. But I've never heard of these rules if they exist, how do people know when making maps what colour the pixel should be? Especially when making new maps, but also when mapping modern external and internal borders?
Edit:
An example of where this matters is the portrayal of microstates. Some people place a coloured tile surrounded by four border tiles, but that makes no sense. Most microstates should simply be a single border tile, or two if they lie across the pixel line, or four (in a square) if they're on the corner. They should have no internal colour.
Maps that would give them a colour would need more than 200 (maybe 700) million pixels [20000 by 10000] as there are 200 million square miles on earth and the Vatican is 0.29 square miles,
Edit 2: I tried to line up the Netherlands border from 2019 QBAM to see if there was any coherent pattern:
I spent ages adjusting the scale and moving it and this was about the best I could get. There seems to me no coherence here. You don't expect it to be pixel perfect but that there is a gap at all instead of an enclave seems really difficult to believe from the relative position of everything else. If you thought well at least I know from this map that all the territory within this area is netherlands you'd be wrong on that, but it seems to me the whole point of a border tile is for when both countries have territory in the same pixel. If something is coloured orange for the Netherlands shouldn't it be completely Dutch?
Edit 3:
This is also very important when discussing sea level rise and fall, or changes to rivers, reservoirs, lakes (Aral, Caspian, etc). At what point will the pixels in the Pacific be changed from land to water? When the islands are fully submerged?
Edit 4: An interesting way of thinking about this problem. If you had to map the earth with 1 pixel, what colour would it be? What about 2 pixels?
Last edited: