I agree. Nothing beats pure text files. These days though, I save them to my Dropbox folder so they're accessible anywhere. I feel like I'm getting the best of both worlds this way.
I have a few three-line scripts in my ~/bin that commit my Org files to a git repo and push to my remote server, or pull changes. Lets me have copies on multiple machines and works really well.
Does anyone have more info on the likelihood of this + the possible ways to solve it if it happens? It occurred to me that most of my important files are potentially deletable via Dropbox.
Now if you accidentally delete something and want to recover it, you have 30 days to do so with a free account. With a paid account, your history is kept forever (basically files are never deleted).
What should happen in the case of missing a file deletion until beyond Dropbox's 30-day versioning windows is that you retrieve the file from your automated backups of at least one of the computers that Dropbox is syncing to.
What would you use for these automatic backups? I use backblaze for my whole computer. However, I think they also only keep file changes for 30 days past change date.
I would pay Backblaze double if they kept past versions.
Colin Percival (the guy who maintains it) is also amazing with support - he was responding to my emails on Christmas day!
If I were _properly_ paranoid, I'd also set up an off-site backup drive - perhaps using TahoeLAFS. I figure the chance of Dropbox losing my data at the same time as I lose three bits of electronics at my house is pretty small.
For non OSX users, you can fairly trivially roll-your-own TimeMachine-like snapshot backup process in Linux using rsync, cron, and a little bash - it's not like Apple invented it... It might even be just as simple using Powertools on Windows.
Every platform i've tried it on has worked okay.
Agreed. Dropbox is more of a consumer disaster recovery solution than a backup solution. If your computer hard drive dies, you lose zero data. On the sysadmin side of IT, we distinguish between backups and replication. Backups have retention and are usually slow to recover from if you need to recover large amounts of data. Replication is often used to maintain a "ready to go" copy of the data. If a server dies a horrible death, boot the standy replication partner and we are back online. Dropbox can work this way of you have it installed on two PCs.
If your data is important, don't leave it to any single service to protect it. If you have truly important data, go ahead and use online backup/sync services, but also make your own point-in-time backups to avoid the accidental deletion of a file you only look at once per year. A cheap USB hard drive that holds monthly or quarterly backups, kept in a safe deposit box, will suffice for most people as that third line of defense.
Rule 1 of backups is "don't keep your files and your backups in the same place". Pretty much common sense.
I was setting up my new desktop to dual boot Debian and Windows. In the interest of saving space, I tried to have Dropbox in both systems target the same directory to sync. I'm not precisely sure what happened - I think the folder hadn't finished pulling in all the files before I logged out - but it somehow trashed everything in my lab's shared directory. Thank god for their API and their automatic file versioning, or I would never have gotten out of that pickle.