Has anyone experience with backing up Scrivener projects with Hazel?
Hazel can watch a folder, say, “my projects” and if a file with the extension .scriv is added it can automatically copy it to somewhere else like the iDisk or a DropBox folder.
For this target folder you can set Hazel to automatically archive (i. e. zip) files with a .scriv extension.
I tried this with mockup folders and a single .scriv file and it worked pretty well.
The question is if a file is saved over and over again (by Scrivener’s autosave feature) will it still be copied every time to the target folder? And if so will this and the consecutive zipping slow down the Mac? Can the usage of options like “altered within the last x minutes/hours etc.” or “not altered within etc.” cushion this?
If it works it could be a work around for automatic zip backups—until 2.0 comes up with an undoubtfully more clever solution.
PS: This is all about folder actions which of course can be executed without Hazel.
What I do is use Scrivener’s zip function to handle that aspect, though if you wanted to off-load it to the OS that could probably speed up the writing process; no need to wait for Scrivener to finish. The rules I have set up are:
- [Kind] is [Archive] (obviously substitute for whatever format you are using)
- [Date Last Matched] is before [Date Last Modified]
That last one will ensure that files do not keep matching over and over. As for the actions, I first copy the file to an “Outbox” folder (where lots of stuff like this goes), rename it with a date-stamp, and then send it to the final destination. In this case, I open the file with an FTP droplet that uploads it to my backup server. This stage could obviously be substituted with moving it into the DropBox folder.
The end result is a remote repository of back-ups in conjunction with a local copy, since I use copy instead of move in the first step.
You may or may not want all of these steps because the way Scrivener works now, it already applies a datestamp so you needn’t supply one with Hazel. I’m doing that because, yes, 2.0 will have substantially improved facilities for backups.
So I guess the key thing is, I don’t let Hazel mess with live Scrivener projects. That’s not a good idea because you could end up producing corrupted back-ups, same problem that Time Machine has if it trawls while you are working on a project. I only have Hazel run against stuff that has specifically been backed up—static, unopened and archived copies of the project file.