We've been running various experiments with mcomb to find out how the biggest problems could be overcome.

My experiences so far is that the tiger data by itself is not accurate enough for route planning. It lacks information on one-way roads so it constantly does stuff like tell you to go up freeway offramps. I have also had problems with dead ends that the current system assumes go through. I have also been meaning to ask Kim if there is enough data present to allow for easiest vs fastest vs shortest route selection. I believe the current algorithm tries for shortest where I would generally prefer easiest (stay on freeways as much as possible, etc) or maybe fastest (quickest route assuming you are going the speed limit the whole way).

But, and this is a pretty big but, the tiger data and Kim's nav system is very helpful as a supplement to written instructions. It is great to know what street you are on, what address range you are in, what direction you are heading, and be able to use the moving map as a reference if you sort of know the area. It is enough to get you around with crappy directions or no directions in an area you are sort of familiar with.

Hopefully Kim will be able to pull some more info out of the USGS data or someone will be able to come up with another data source.

Anyway, after using Kim's software for a couple of weeks I am still very impressed. An accurate datasource still seems to be the biggest missing piece.

-Mike