Apr 2010

performance improvement

I've changed the way that Satellites renders the 3-D sphere that forms the basis of the globe. Instead of computing all the vertices of the 6 degree tessellation, I've pre-computed them and put them in a table. The table isn't very large, and the computation involved a lot of trigonometry, so I think the trade-off is OK.

New shorelines ...

Satellites used a fairly old CIA dataset for the outlines of the continents, islands and major inland seas and lakes. This dataset had been cleaned up and re-issued for other uses by NASA, and Satellites used that cleaned version, but I began to notice things missing -- mainly 'small' islands! South Georgia, for example, in the South Atlantic wasn't shown.

I went hunting and found another dataset with more detail, actually a
lot more detail than I wanted. It's lowest resolution was about the same as the highest resolution of what I was previously using. But it included all the missing islands, and a few other things I hadn't noticed. The downside is that it may be too complete!

Here are some comparisons, what do you think?