Niagra

  The Virtual Earth blog has a great demonstration of what can be done with state of the art data and imagery in creating immersive (or, rather, more immersive) digital environments. One thing I've noticed in general with elevation data found in 3D geographic tools (Google Earth, Virtual Earth) is the lack of precision where there are sharpt/vertical precipitations. But this isn't the case with this great example of Niagra Falls in Virtual Earth.

Compare this with, for example, the white cliffs of Dover which, according to photographic evidence, are vertical:

but which appear in Virtual Earth somewhat sloping.

One thing is certain - digital representations will steadily continue their march from approximations to photorealism. The only question is how rapidly this will happen. On that note, the more crowd sourcing that can be involved, the quicker we will half-life the planet.

Comments

[Raj Rao]

There is a reason for this lack of vertical precision. Its caused by the interpolation of the elevation points that were used to generate the elevation raster used to give the image a 3D elevation.
http://therajahs.blogspot.com/2007/07/3d-maps-lack-precision-in-google-earth.html

[Amit Patel]

3d maps are usually represented as an "elevation" for each latitude/longitude; it's more efficient than the alternative representations. The most common wart is the one you show, of vertical surfaces. However, even if vertical surfaces are handled specially (by storing a "discontinuity" in the underlying data), we can't represent overhangs, areas underneath overpasses, tunnels, caves, etc., by using a single elevation per latitude/longitude.