OpenStreetMap
I wasn't terribly surprised by the fragmented roads imported from TIGER. Identical fragmentation was evident in other maps. It's great to have a starting point. Really great. Wouldn't be this far along now if I had to trace every road.
Some roads are worse than others. Yesterday I tackled the Telfair County Line Rd / County Line Rd. This is a highway that meanders across the Dodge and Telfair county line. Both counties share the expense of maintenance. However, for what ever reason, this road does not share the same name. This is unfortunately common where roads cross from one county into the other, but in this instance it's downright aggravating. All sections in Telfair are called County Line Rd. All sections in Dodge, Telfair County Line Rd. Sometimes the sections are very close together. One part weaves over and under the county line, and over relatively flat land, too.
It's going to take more than trivial renaming to fix. I'll have to consult more than a field inspection to figure it out. But no one said this was going to be simple when I registered.
A few roads are blasts from the past. I was drawing in a new road in Little Ocmulgee State Park, and noticed the County Road immediately east. Except it's not a county road anymore. It's an access road that semi-follows the route shown on OpenStreetMap. But the County Road number gave me pause, and I gave some of the other roads closer inspection. Some look as though TIGER and the USGS drank from the same trough. But others, like the county road, may predate the park itself. Little Ocmulgee State Park was built by the Civilian Conservation Corp in the 1930s, so I suspect some of these roads were mapped before then.
Editing these fossils gives me an odd feeling, almost like desecration. They'll be no record of their existence after they're edited. On the other hand, maps, by definition, are products of their time, and keeping non-existent roads is just a source of confusion.
Some, though, seem to be outright phantoms. The road immediately east of US 441, south of the airport, never existed as far as I can tell. It would have run right over a steep sand ridge. Granted it might not have been so steep on the south side before the state took soil to build the adjacent overpass There's no sign of an old road in that location on the site, not even a utility ROW that could have been mistaken for a road. Where TIGER came up with that one . . . shrug.
Right now I'm tinkering with JOSM and MerKaartor, to see which I prefer, to expedite map editing. Potlatch is all right, but editing maps offline may be faster on my computer.
Discussion
Comment from ae6mb on 23 January 2011 at 19:29
If it makes you feel any better, OSM saves the edit history :-) That means the old TIGER is still hiding back there. I have to say that I feel good when I fix some screw up from TIGER. To me, it is the satisfaction you get when you fix up a broken down old car or something. Then again, I might just dislike TIGER from having to use it when I was a Census enumerator. You can really end up in some weird places following what TIGER says are roads.
As for TIGEr predating the 1930s, I think it was made in the 1980s, but it was based off data made in the 1960s, which was probably based on USGS maps and census worker data or something. In theory the census maps are updated every decade, but I know people who were part of doing it a few years ago, and a large portion of their updates just disappeared once they got to the office, and never made it on to the maps we were using, as they should have. I thing USGS is a lot better and more professional at mapping than the Census bureau, but they, unfortunately, don't have the budget of the Census, so they can't update their maps nearly as often.