Hello, I think that CouchDB is a good fit for the project i'm working on but I'm not sure. I would appreciate advice. what I have is a central server in charge of tracking items, their bar-codes and whether those bar-codes were scanned and several devices doing the scanning. performance is important. data has no relations.
due to customer requirements I need the data on the items locally on the devices and as current as possible. the network connection is expected to be patchy at times. when a device scans an item it must be notified if it was already scanned. what I thought I can do was: 1) run a local couchDB server on each of the devices and have them continually replicate the server's database, this way I have a good syncing implementation that I don't need to build from scratch. 2) to keep things simple at this point I will only make actual changes to the documents on the server and have those changes replicated onto the clients, trusting couchDB to send me an error HTTP response if two devices try to mark the same item as scanned (the first will update the document, then the second's command revision won't be appropriate anymore) since for now I'd rather avoid conflicts than deal with them. I have come across criticism that this will leave me very susceptible to network errors, is the incremental replication process give me reasonable protection? does this sound workable? has this issue been resolved yet? => http://www.mail-archive.com/[email protected]/msg02711.html
