While work on TopOSM (http://toposm.com) has been somewhat slow recently for various reasons, there has been some important progress that I wanted to share.
Currently I'm attempting to create a TopOSM-style map for the entire United States. While that may seem almost trivial, since two states are already finished, I can assure you that it is not.
The first step, which is, hopefully, complete, has been to gather all of the required data. In addition to OSM data for most map features, I've been using the USGS National Elevation Dataset to generate hillshading and contour lines. I recently received the approximately 600 GB large 1 and 1/3 arc-second NED data for the entire United States, plus several other interesting datasets. Additionally, I have received the National Hydrography Dataset - a detailed dataset containing rivers, lakes, wetlands and other hydrographic features - for the entire United States. The latter is (slowly) being imported into OSM, by the way. I really owe the helpful people at the USGS a big thanks for providing me with these!
The fact that these data sets are very large make them at least an order of magnitude more difficult to work with than what I've previously encountered. Things that should work just don't. There are countless cases where large data files have to be split and processed sequentially because of memory limitations, disk space or just plain bugs or limitations in the software. Once everything works, the project will require a massive amount of CPU time and disk space to render. That's a later problem, however, and it's a job that can be distributed if necessary.
At this point, I have an almost working set of scripts to process all of these different data sources and finally render map tiles. There are plenty of bugs left to work out and many needed improvements, but at least it's moving forward. I'll post further updates and examples here as the work progresses.