mmcirvin

What, no mention of the Fuller Projection? ;o)

There's a page there on polyhedral projections, of which Fuller's is one, though the idea predates him.

Notice how most of the examples there have Kamchatka on the left. I wasn't making it up after all!

Yeah, I think in those cases it's the result of trying to get South America nicely centered on a face of the polyhedron, since South America is where the page's author lives.

The same thing happens if you try to do a simple interrupted map with New World and Old World hemispheres, which used to be very common. Dividing Far Eastern Russia like that may, more than anything else, be a consequence of trying to show the Americas and the Old World in an aesthetically balanced way.

(I once knew a guy who was from that part of Russia. Americans often refer to it as part of Siberia but I think it's actually east of Siberia.)

See also the star projections, which, like Fuller's, concentrate on showing proper polar relationships between Northern Hemisphere continents while dividing up the Southern.

I've always thought it interesting that, while world maps are a difficult problem and the choice of projection is fraught with controversy, for smaller-scale maps the issue is almost completely settled. Almost certainly, the most popular map projection in the world today is the 18th-century Lambert conformal conic, which is essentially never used for world maps but is the usual projection of choice for smaller regions in the temperate zones (you see it in most wall maps of the contiguous US, for instance; if the western border with Canada looks curved, chances are it's the Lambert conformal conic). For such a purpose, it doesn't distort distances and sizes enough to really matter much, and it's conformal so everything is locally the right shape.

For regions near the Equator or with a larger north-south extent, various other choices are more useful, like the polyconic, the Mercator or the transverse Mercator.

I have a Hammond atlas from the mid-90s that did something interesting for its continent maps: they came up with a computer program that could generate what they called an "optimal conformal projection" for any large portion of the Earth. They'd draw a region bounded by smooth curves, and the program would provide the conformal mapping that minimized scale distortion within that region, by some measure that they didn't specify (probably something like RMS scale variation integrated over the region). The maps looked pretty nice, though in most cases I'm not sure they were really that much better than a judicious choice of one of the old standbys.