Piotr wrote: ↑Sun Apr 05, 2020 11:40 pm
Do you know that because you have tested it? What is the source code?
I know because I experimented with it a long time ago and have experimented with many conformal projections constructed via complex polynomials.
Snyder shows what happens if you render even a little ways outside of the useful boundaries. Notice the lower right corner, and notice the abrupt bend in the lower parallel toward the west edge that results in a large expansion of that region.
GS50, slightly expanded perimeter
IMG_0275.jpg (62.7 KiB) Viewed 17868 times
He writes,
It should be stressed that most transformations resulting from the use of [these equations] can only be used within a limited range, depending on the number and value of coefficients. As the distance from the projection center increases, meridians, parallels, and shorelines begin to exhibit loops, overlapping, and other undesirable curves. A world map using the GS50 projection is nearly illegible, with the meridians and parallels intertwined like wild vines.
It should be stressed that most transformations resulting from the use of [these equations] can only be used within a limited range, depending on the number and value of coefficients. As the distance from the projection center increases, meridians, parallels, and shorelines begin to exhibit loops, overlapping, and other undesirable curves. A world map using the GS50 projection is nearly illegible, with the meridians and parallels intertwined like wild vines.
— daan
What if you instead took the overlapping sections and separated them (for example, in Orthographic, the world would be in 2 pieces), so the map is in multiple pieces? What would that look like?
What if you instead used an interruption scheme to render the world?