Stafford Springs

Posted by Tomash Pilshchik on 28 October 2009 in English (English)

Yesterday I did a lot of cleanup in Stafford Springs. I mainly aligned the TIGER tracks with the Yahoo aerial photos, but I also used some of my GPS tracks for reference.

One thing bothers me a little. I notice that a lot of town boundary shapes have been imported. Some of these appear to run along roads or rivers. But, what if they are in the wrong place? Do the TIGER shape files define the town boundaries, or are they an attempt to represent them? If it runs down the middle of a road and me 'move' the road, should we move the town line too?

Also, in Merkaator, it is too easy to move a way by accident. It is almost impossible to fix up roads near a town boundary without accidentally grabbing the boundary and moving it. Then it is not clear how one moves it back to its original location. Any suggestions?

Location: Brendon Street, Stafford Springs, Tolland County, Connecticut, 06076, United States of America

Comment from Ethan O'Connor on 29 October 2009 at 19:44

Tomash -- generally if the TIGER data shows a boundary along a road or stream, the boundary falls in the middle of that road or stream _as they exist in reality_, not as they exist in the TIGER database. Perhaps the best approach to to add the segment of the road or stream that the the boundary follows to a relation defining the boundary. This way you don't have to worry about pinning the two ways together, and as the road or stream gets updated to a (presumably more accurate) alignment, the boundary will get the upgrades as well.

There's active discussion about whether to do this with a relation of type=boundary or as a type=multipolygon with a boundary=administrative key:

After quite a bit of consideration, I believe I come down on the type=multipolygon side of the discussion, but in either case it's better than trying to keep boundaries and the underlying ways the follow synchronized!

Login to leave a comment