Yup, understood. The point you raise was part of the problem. That is, I wanted to keep using my public tree, because people have started finding it and using it, but the approach you describe takes my tree away from them every time the synchronization fails like it did for me. They might or might not find each new tree, and if they do, they might not realize it's just a new version of a source they've used before.
The alternative - downloading from the tree - gave me a different problem: loss of information.