(In reply to comment #7) > > Can you name a specific use case where you expect this to be beneficial? > > An use case based on some feedback I received is: > Stage 1) Sync NAME,TEL,EMAIL > - First time DB population it will quickly create a browseable DB > the fields chosen are the most important required for the > system usage. > Stage 2,3,4,...) Sync (Other Important Fields) > - The DB is populated with extra (less important) fields > that may be required by the end user but it is nice to have now. What is the expected benefit of doing stage 2,3,4 etc separately? "More quickly" is probably the answer, but how quickly does it have to be to outweigh the downside (longer time until the cache is really in sync with the peer)? I suspect that whoever asks for this overestimates the time that it takes to download those extra fields. Before we implement this, someone should quantify that by running some download experiments. You can use SyncEvolution for that, see the src/pbap/README for instructions on how to filter fields.