It's interesting from an artistic perspective (as was the parent discussion) in that it shows the shortest path from a place to a central location as well as the most-traveled ways into that central location. The original site references has some fun examples that are colored, as well. I'm just not so sure that it's answering an interesting technical question. Not for the sorts of things that I'm interested in, anyhow. This operation requires a preexisting transport network and identification of a central place. Probably just a lack of imagination on my part.

I'm much more into creating mapping stuff like rivers and roads (totally outside the context of this sort of thing, which focuses solely on traversing existing graphs). For that sort of thing, something like A* would probably be a better tool because for creating the paths in the first place, you're most likely to be traversing a gridded cost surface than an existing network. Start with flow paths for water, then flow paths for goods and people. Then population growth and so on. Top it all off with the ability to force things (drastic increase in cost function in an area due to drought or a new xenophobic race or decrease in cost due to adding a new lake or river) and you're at the classic god game. That's why we're going this, after all. Or maybe it's just me...