Thoughts On Google Earth

Today I observed Kate and Galen solving a geography puzzle using Google Earth. It brought to mind a couple ideas about Google Earth and how it is changing the way we see things.

The problem that Kate and Galen were working on was to go from Isfahan, Iran, “northeast, and find a capital city below the Hindu Kush, on the banks of a river of the same name, west of the Khyber Pass.” They were using Google Earth and they were having trouble doing it.

fig. 1: National Geographic's map of Central Afghanistan

This is first because Google Earth, contrary to popular impression, is not a map. It is a photograph. A map contains interpretation whereas a photograph is raw data. On a map, certain things are pulled forward, such a capital cities, and other things are pushed back (small villages). Collective features are named, such as mountain ranges and deserts. Linear features that are actually very thin, such as rivers, are named and marked. The interpretation offers you a reading of the landscape where some things are important, and others are not. A detail of National Geographic’s Asia map (fig. 1) illustrates map-making.

Google Earth in contrast offers everything in equal weight. There is little interpretation-–pretty much limited to whether there are GE icons there. Note that a GE “map” of Afghanistan (fig. 2)

Fig. 2: Google Earth, Afghanistan

is strong on point and polygon features (towns, countries) because its easy to give the computer automated ways to place labels for these. It’s correspondingly weak on linear features such as rivers, and features that don’t really have boundaries, such as passes and mountain ranges.

The reason is that there’s no easy way to tell a computer how to label such features–where do you put the river’s name along the river, and how many times? How do you arc words to suggest the imprecise yet very real bounds of the mountain range? It’s this kind of labelling that human cartographers do well: each label is put on with love (which is why it’s unwise to rush a cartographer, or the love disappears).

The answer is Kabul, Afghanistan. It’s so easy to see this on the Nat Geo map, which identifies the Kyber Pass and the Hindu Kush, and so hard on the GE photo, which does not. At first it’s tempting to say that the Nat Geo map is therefore better. However, I’d suggest that the reason the quiz question was phrased the way it was is because we are used to looking at maps, not photos.

GE photos are strong on point features and coordinates, so you wouldn’t describe a place as “below the Hindu Kush, on the banks of a river of the same name, west of the Khyber Pass.” You’d say it was at latitude 34.5541N, longitude 69.1594E, elevation 1890m above sea level. Also, fFrom a quiz point of view there’s not much fun in that.