I don’t believe I’ve ever heard anyone talk about it before, so I can only offer general advice for evaluating whether a sync server is going to work well for you:
Firstly, there is the matter of whether one needs a heavy-duty “always synced eveywhere” workflow or not. If you just intend to use it as an offsite location for your automatic backup folder, or intend to use alternative, safer methods of data transfer, then it’s almost impossible to find something that doesn’t work well with Scrivener that’s a serious product and not some early tech demo or something. Because at that point we’re talking about copying .zip files around, and anything should be able to copy one file a few times a day.
For those that use multiple computers and want the much more complicated process involved in keeping a “single” project up to date on multiple devices for the sake of convenience, this series of tests should help expose any rough edges or flaws:
-
Create a disposable project that you don’t care about. The Tutorial in the help menu is great for this, as it has a lot of material in it and it doesn’t matter if it gets scrambled as you can just make a new one.
-
The next step requires knowledge of how the specific system works, so there is no simple instruction for saving your project in such a way that it picks it up. Whatever it takes, the idea is to save the project in a location where it syncs constantly. That may be one single massive folder, like how Dropbox works, or you may have to set up sync folders from existing folders on your system. The system I use (Tresorit) requires you to pick what syncs from your existing files. So in my case I would create a “Scrivener Test” folder somewhere, set it up to sync, and then save the tutorial project into it.
-
At this point you should have the test project open. Just do anything, like adding a new item to the binder and typing in a word or two. Look around on your screen. Is it obvious that sync is working? If not, check the task tray and make sure its icon is set to always show, and try again.
Some sync tools do a poor job of communicating when they are working, and when they are done—when it is safe to shut down or sleep the computer and switch devices. It’s not a deal-breaker, plenty of people use iCloud for instance, but it does require extra vigilance and learning how to get that information somehow. Maybe it has a client window you need to keep open all the time and monitor at the end of the day. Maybe it embeds feedback in the file manager and you have to keep your project folder visible.
-
If convenient, switch to the other machine and locate the project in the sync folder. Try to open it. You should get a warning about it already being open on another machine. If you don’t, that’s not a good sign, that means even though both devices show they are up to date… they aren’t. Cancel loading if applicable, but if it let you open it without fuss, then you probably already have a mess and will want to start over with a fresh tutorial, or maybe evaluate whether it is worth it to continue testing.
I’d be nervous about the service at this point, but the first thing you’ll want to check is through the service’s settings, to make sure it doesn’t have a bad default that doesn’t actually keep your computers synchronised without manual setup. See below for more on that.
-
Close the project on the first machine, wait for it to sync, then on the other machine let it sync and try again. It should open without any hassle. Are your edits there?
Go back and forth like this for a bit, opening and closing, making changes, do some large-scale changes like changing the default font, selecting everything in the binder and using
Documents ▸ Convert ▸ Text to Default Formatting, followed immediately by another bulk change while it is syncing, and make sure that all of those successive changes made it over the other machine just fine. -
Next, check what happens when things go wrong. Unless you are perfect about making sure sync is always done before switching things off or what have you, one day you’ll accidentally end up editing the project twice from two different locations, causing conflicts. How a service reacts to that is thus good to know about.
To test that, pause sync on one of the machines and open the project in both places. Make some edits, some deliberately to the same items in the binder.
Close both, turn syncing back on, let it fully update, and then open either of the copies. Scrivener is designed to detect the most common method services take to handle a confusing situation like this: duplicating each file that conflicts. If that is how it works, you should get a warning about such files being recovered, and copies of them placed in the binder. Make sure all the files you intentionally messed up are there.
If that doesn’t happen, it may be worth investigating further how it handles conflicts.
If you run into numerous problems that seem to never resolve, double-check the service does not have any kind of “smart sync” feature, and if it does, that it is turned off on all machines, and try again.
For iOS, you’re going to be using a procedure much more like the alternative linked to above. You’ll be fetching the latest .zip backup from the synced backup folder and using Files.app to copy the zip into Scrivener’s storage, and unzipping it. When you’re done, you would zip the changes back up and copy them into your backup folder, then replace your desktop copy with that. It’s a few extra steps, but it’s not that bad once you get used to it—it’s all I use.
With iOS in the equation, it’s worth considering not “live syncing” at all, because it’s always a bad idea to put both your automatic backups and your live projects into the same service’s realm. You want to avoid putting all your eggs in one basket, and with iOS being unable to provide sync services to software like a desktop can, you have to cater to the lowest common denominator.
Whether to mix everything together is a judgement call of course. It depends a bit on the service as well, and the context. If you’re going to be sharing files with someone else for example, then you have to account for twice the amount of potential user error.