Increased reliance on optimistic data replication has led to burgeoning interest in tools and frameworks for synchronizing disconnected updates to replicated data. We have implemented a generic, synchronization framework, called Harmony, that can be instantiated to yield state-based synchronizers for a wide variety of tree-structured data formats. A novel feature of this framework is that the synchronization process—in particular, the recognition of situations where changes are in conflict—is driven by the schema of the structures being synchronized. We formalize Harmony’s synchronization algorithm, prove that it obeys a simple and intuitive specification, and illustrate how it can be used to synchronize a variety of specific forms of application data—sets, records, tuples, and relations.