>We use optimistic table buffering. At least I think that's what's being use for this. Anyway, it seems that somewhere in our code something is happening where the locking is not working right. We are not sure where yet, but it looks like we have a valid Record-level lock, but another user can change the same record. What are some of the things we can look for? I was wondering if it was flushing (where is Jim when you need him?) but don't know enough about that. Any guidance appreciated.

IIRC, record locking only occurs when a TableUpdate() has been issued, therefore, someone else can edit the record and save changes before you do. One way around this is to use GetFldState() or use the nUpdateWhat param in your TableUpdate