ndn-sci

The attachment is the UML diagram for PublishAdapter (the current implementation for the publication process).

I use PublishAdapter for the sync update, because

• the ChronoSync requires callback function to process the updates and validator when applications use the ChronoSync library, and the callback's job is to decide whether to fetch the listed missing data, I think a function is good enough to do the job;
• the publishAdapter is tightly associated with the "sync processor", e.g., when the PublishAdapter fetches the data and verify it, it injects the data into the local DB, and then needs to notify other sync nodes by publishing it over ChronoSync;
• for the Sync Update Executor, I'm not sure if it should be a separate class, I think its job is the data insertion, and the publishAdapter contains the data insertion as well.

Correct me if I'm wrong.