3D maps lack precision in google earth and virtual earth

In a recent post by Mathew Hurst, he wrote about how there is a lack of precision vertically in the 3D Earth models (http://datamining.typepad.com/data_mining/2007/07/niagra.html). There is a reason for this. Most of the elevation data is sourced from point clouds that were either generated by a scanner aboard the Space Shutter (Shuttle Radar Topography Mission or SRTM) or by a laser scanner aboard a plane (LIDAR). Typically, most of the data is from SRTM and so the points are spaced far apart on the ground. These points are then converted to a digital elevation model via a process of interpolation. So consider what would happen along cliffs.... One point would be on the higher side and the other on the lower side. Interpolation algorithms dont know that there is a cliff - so they generate a smooth surface from the high side to the low side. In the above images (shown on Michael Hurst's site) - the bottom image is from Virtual Earth and appears sloping because of this problem caused by interpolation. In the field of GIS (the side where such errors might not be acceptable), analysts introduce break lines to the interpolator. Points on one side of the interpolator do not influence points on the other side - thereby mainitaining a nice cliff. In Michael's post - he also shows another screen shot of the Niagra falls, where the falls appear nice and vertical. Such an effect could have been achieved manually, or by the use of break lines - as described above, or terrestrial LIDAR scanner could have been used to get a much more denser point cloud to accurately model the waterfalls. As automated techniques are developed to generate break lines, as well as higher resolution elevation data is obtained for the world - more of the Earth will begin to look photo-realistic in these 3D environments. (Many government organizations have such data already for their states/counties - and I think that they might start providing their data - or even start creating such models - for use in common 3D GIS platforms such as Google Earth and Virtual Earth