I got turned onto a tool called Hootenanny (GPLv3), which is for
finding and fixing conflation errors. Right now it only runs on Centos
6.7 or Ubuntu 14.04. Now that I have it running in a VM, I'll try to
upload my VM image so others don't have to go through the pain of
getting it built and installed, which was pretty painful. The project is
at: https://github.com/ngageoint/hootenanny/blob/master/README.md. I've
managed to get it to work offline.
Basically it can read in Shapefiles or OSM files, and let you merge
them together and export an OSM file. It uses a modified version of the
ID editor. This is great when you have multiple datasets for an area
from multiple people and you need to clean it up and merge it all together.
I'll be merging two large datasets together with a lot of overlap,
over the next few weeks, I'll let everyone know how the experience goes...