Hi, I wanted to confirm whether autosave files were saved in folders based on time (the version saved at 9:00pm is in a folder, and the version autosaved at 9:10 is in a folder, up to a certain number of versions), or based on the data (Chapter 14 and 15 are being saved every five minutes to the same location).
The reason I ask is that while I mostly work on Scrivener from home, I also do some writing when I have spare time at work, so I’ve copied the scrivener files back and forth from a flash drive that I can carry to work (I may start using dropbox instead, but as I can’t download the dropbox client at work, the flash drive seemed to be the better solution).
To make sure I’m still working on the same version of the file, at the beginning of my home writing session, I’ve been copying the files from the flash drive to my hard drive (both because working on the copy on the hard drive seems to work faster than working directly off the hard drive, and to make sure the manuscript is backed up to my computer) using Allway Sync Pro, and I use the same program to sync the files back to the flash drive after I’m done writing.
Now for some reason (and this is an issue with the other program, not Scrivener), Allway Sync seems to exclude about 1-16 files from the Scrivener folder from the sync. So in order to sync those files, I have to manually tell the program which file to sync in which direction (normally, it copies the most updated version over the older version).
Last night, I wrote until I was about half asleep, and decided to make sure I saved my progress to both places before going to bed. Except, for some reason, when it came time to tell allway sync what direction to go in, I accidentally told it to copy the old version over the new version, and did not realize what I’d done until I had already synchronized the files. It looks like I lost about 700 words. As I said, I realize this is not an issue with Scrivener, but I figured it couldn’t hurt to find out if maybe Scrivener might have managed to save a version of the file (the thought being that if it was saved to a new file, for which there was no matching one in the old version, maybe it wouldn’t have been copied over).
I figure it’s a long shot, but worth asking. Thanks again for a great piece of software!
Ah, no, sorry to say when it auto-saves it does so right over the old copy of the file; it would waste an awful lot of disk space to act otherwise. So if you are writing in one portion of your outline, say a chapter with five scene files in it, then you are going to be hitting at least five RTF files with changes. You might also impact their linking and notes files, and there will also be some project-level control files you’re changing as well, such as the search index and the master binder control file. Who knows how many auto-save events happen in a typical writing session, but it’s probably in the hundreds—so you can see why it wouldn’t be feasible to stack files up.
Hopefully that sheds a little light on why your synchronisation software isn’t moving the whole project over, either. Only the files that are altered get saved, the vast majority of a project, in many cases, never even gets opened once by Scrivener in a typical session. So you should be able to trust your sync tool. I’ve used a similar tool in the past, and it worked well for me. Do keep daily backups though. Any time you or a computer program not named Scrivener is messing with the internal components of a project is a higher level of risk (and that includes Dropbox). We’ll be making that last part easier for you, soon.
It doesn’t entirely answer the issue with the other program, since it’s specifically excluding the files from the sync, not saying that the file hasn’t changed (it has separate sections for new files, deleted files, changed files, unchanged files, questionable files (like if one version is newer, but the other has more information), and specifically excluded files, several of which are temporary files. I’ve noticed in the last couple of days that it’s also been excluding the excel spreadsheet I’ve been using to track all my nano stats, which it never had a problem with before, so I’m not sure why it’s excluding it. Ive opened up a support ticket with them a few days ago (I suspected something like this might happen when I noticed the exclusions, and was hoping to find a fix to avoid it before it happened), but haven’t gotten any answer yet.
Anyway, thanks for letting me know! Time for me to try to rewrite what I lost.
UPDATE: Well, I figured out what was wrong wit the other program. It seems I had an exclusion filter set up for changed files. Why, or how, I don’t know, because I don’t remember doing it, and since the changed files are what I’d be trying to sync, I wouldn’t want to exclude them, but I took it off, and analyzed my files again, and it looks like it checked everything. So hopefully, I won’t run into this issue again
Sorry, I don’t agree with the awful lot of disk space. My novel is still under 1MB and if I made new copy every day, that would make it 30MB a month and 365 MB a year. Today’s computers can handle a lot more than that…
I am having problem with autosave also. What if I don’t want to overwrite what I just did? I would like to have control over what and when I save. Also it would be nice a question at the end asking if I want to save. That way I would be sure I have saved and won’t have to be paranoid about double checking:)
I am not very happy about the automatic backup options either. If I set it to backup at close it makes a backup copy even if I just open the file but make no changes.
It should make a backup only when there are changes to save. But I understand that there are always some files that are changed, even if I just open it to read for some reason…
I apologize if I misunderstood something. I am pretty new to scrivener, although I have a Mac version and used it sporadically.
A new copy every day? That’s what automatic backups do; give or take. While we do rotate those out at five by default, you can turn that off and have your 365mb. I’m not sure what this has to do with auto-save though? If you work all day in a project that’s way more than one copy getting made of it, and you’ve got to consider that a lot of folk are putting way more into the program than you have. Some are writing whole series in a project because they like having all of that common universe data all together. And there is the matter of those that work in smaller pieces. A typical 150k word book in 600 outline nodes is represented by (with average use of the meta-data features) roughly 1,500 individual files on the hard disk just to represent its current working state. Let’s say a typical session is three hours in a day. Auto-save isn’t going to be firing every two seconds of those three hours in most cases, but it won’t be too far off from that. Let’s say on average every ten to fifteen seconds. That’s about how long it takes to compose a sentence or two, and let’s be generous and say our author only pauses for a moment then and while pausing auto-save saves the work. So in a 3 hour session that 3,600 seconds which gives us lets say 288 auto-save events. That’s probably conservative. Each event might write two to four files, so one working session produces let’s say 900 new files. That’s a part-time writer spending a few hours a day in a book. Even if they only work five days a week on average, that is roughly 250,000 new files in the project every year. There are going to be people taking Scrivener way further than that, producing millions of files a year across a dozen projects.
That, is not an acceptable toll, even 250,000 files is ridiculous. It would take a very long time just to make a zip backup of that. Computer performance isn’t just sheer mb, but file quantity based. Lots of tiny files that add up to 20mb in size take a lot longer to process than four or five files that add up to 20mb. Consumer computers aren’t design for those kinds of workloads. Unless you’ve got a closet full of rack-mounts and RAIDs and an $800/mo power bill, you don’t want million file projects. When I said “disk space” I meant the total equation. The I/O performance profile, the bulk, the management load of keeping track of hundreds of thousands of files. It would mean having to switch out nearly every piece of technology in Scrivener and moving to more robust but less stable mechanisms like relational databases that require regular maintenance and optimisation.
And for what? 298,500 micro change files you don’t and won’t ever need?
From the user manual’s quick tour introduction, pg. 29:
The Implications of Auto-Save: Whenever you pause for two seconds, Scrivener’s
auto-save feature will kick in and save all your recent work. This is a very key thing
to be aware of, because many people have cultivated the habit of relying upon their
last save point as being a bit of an anchor point. They can write freely and in an
experimental fashion, and if what they’ve done displeases them they can just revert
to the saved copy on the disk. The good news is that Scrivener is designed to work
with this method, and in fact its way of doing so is in general superior to using the
traditional model, because whenever you put down an anchor point, or a Snapshot, in
Scrivener, you leave a trail of them. You can view them all in a list later on, preview
their text, and roll back to any point in a document’s history. This is a little out of
scope for this quick tour, so for now put it on your list of things to look into if this
is how you work, and when you’re ready to start writing, check out Using Snapshots
You’ll get used to working in a safer system in time. I never even think about what is saved and what isn’t when using Scrivener. It’s all just always exactly what I tell it to be. No mental tasks and checklists to take care of; just be. And this hasn’t hurt me in other programs either. I still have a reflex to hit Ctrl-S in other programs where a crash or mistake can mean hours of lost work otherwise. In other programs I’m always worrying about that kind of stuff, and having to make decisions about whether I should save now or later.
Scrivener obliterates all of that. The computer handles the mechanical junk, and you handle the strategic decisions with your list of snapshots. You don’t have to worry about maybe saving or maybe not saving because you aren’t sure about the direction things are going it. Just fire off a snapshot anyway; maybe even name it something reflecting that, and continue onward. Later if you hate it you can toss the snapshot and roll back to your last save point. Or you can roll back and keep the snapshot anyway. Maybe the overall was bad but there were a few good lines you’d like to keep around.
See, to do all of that with a .doc file you’d have to messing with file names and opening and closing a bunch of files wondering which is what to sew it all back together. We don’t think authors should have to waste their time with all of that secretarial work.
Yeah, that’s the problem; and it is hard for Scrivener to figure out what you meant to be doing in a way that is safely always going to be protecting other people who work other ways.
If this becomes a problem for you, either boost how many backups get saved, or turn off the open/close stuff and leave it so manual Ctrl-S events fire a backup (just don’t reflexively hit Ctrl-S all day
Recently, Keith added a check to the Mac (beta) version that if NOTHING is changed by the user (i.e. nothing triggers an auto-save event), then the automatic backup is not triggered on close. I’m betting when things settle down, Lee can add that behavior to the Windows version. I know autosave’s current behavior seems wasteful, but if it’s really that important to you not to have a redundant copy, just delete the backup that you accidentally caused.
As Ioa just outlined, the autosave feature may be an adjustment, but it’s really far better than the alternative. I’ve never lost work to a power outage or a reflexive click on the “Cancel” button instead of the “Save” button when closing Scrivener, but I have lost work to both situations with other programs.
Also, if you have an automatic backup from the last time you opened Scrivener, you can always get back to what you had before you edited your work. It may be more tedious than hitting the Undo button a bunch of times, but you still have that version of your writing if you forget to use the snapshot feature.