In the recent review of online mapping sites, one of the comments about Google Maps was that it was missing some key datasets, notably minor water features. However, if one looks at any of the many Google Maps mashups, one can see that there is alot more data available than just roads. Compare these two images of the same area. The top one is taken from Google Maps and shows only roads and some railways. The bottom is taken from a Google Maps mashup and shows roads, more railways, parks, minor water features and commercial/industrial features. As well, the data seems to come from different sources (Navteq for Google Maps versus TeleAtlas for the mashups), resulting in slightly different geometries and road classifications (it would appear that the Navteq version is more correct). Why the difference?