srfreeman: I really don't know what the FSCK you are talking about. ownCloud was most definately build to be a product to feature everything dropbox does - and more. No offence, but this is absolute nonsense:
BTW; If you get into development and would like to add functionallity to make the sync portion of ownCloud more like the Dropbox product, look into the other open source project librsync.
Things are rather the other way around. ownCloud was and is and always has aimed to be able to replace dropbox for normal home users as much as possible and we go through extreme lengths to make it as easy as possible. That is why it is written in PHP in the first place (runs EVERYWHERE), that is why it still defaults to SQLite despite its performance issues (pissing off many advanced users
!) and why ease of use is one of the central development goals.
If you have ever sat together with Frank Karlitschek (our benevolent dictator
) and Jan-c Borchart (lead designer), you would know that the day ownCloud ceases to be useful for normal home users and exclusively targets enterprise usage is the day they will fork ownCloud to bring it back to its original goal: free the data of NORMAL home users.
@jblack: I can guarantee you that ANY improvement to the performance of either the client or the server, or their interaction, would be more than appreciated. If you don't believe it, I hereby invite you to come to the ownCloud Contributor Conference in the last week of August in Berlin (which I am organizing) to work directly with Klaas and Daniel and the others on the client and/or server.
I hope you have tried ownCloud 7 and have noticed we worked real hard to make it faster in many ways. As Randolph said, the server part of using change notifications is afaik now also implemented and that should help keep cpu usage down. And I believe we now do parallel, chunked uploads - still not 'only changes', we didn't get to that but help is welcome. Also - it still kicks off a full sync ever 5 minutes, I've asked why that is:
ogoffart: the problem is that changes can happen durring the sync
ogoffart: and we disabled the watcher during the sync
ogoffart: for the obvious reason that the sync is modifying all the files
jospoortvliet: ok so you have to re-scan everything after the sync? sync modifies ALL files?
ogoffart: Well, all the files that it downloads
ogoffart: + the journal
jospoortvliet: just for my info: is it possible to do this better and did we simply not have time to do this, or is this the best we'll ever do?
ogoffart: the problem is also that the watcher does not tell *which files* were changed, just the directories in which something was changed
ogoffart: When the watcher tells us exactly which file change, it is possible to ignore the journal or file that we are currently syncing
So, this would need work on the watcher. You say you have Qt and C++ coding skills - if you're interested in helping to fix the issues you have I'm certain that would be welcomed with open arms. Ping me if you need any help - google should give you a mail address