Auto-save

Yes, I’m agreeing with you. I also don’t have much knowledge of how this kind of thing actually works, however. At least not on the PC.

After reading through the whole thread, here’s what it seems the OP is asking for:

OP likes auto-saves but is of the opinion that, like “Windows norm” (aka Word, etc) on opening the program (1) there should be a temporary “file” created where these auto-saves occur (for Scrivener, of course, it would be a temporary project directory). (2) Manually saving would save to the “master” project directory (I think it was called “working” directory; “only the user should be able to save there”). Then the OP thinks backups is great, but doesn’t like the rolling auto-backups. (3) OP thinks the backups should be manual backup from known “good point”, that the user chooses as something like “complete”.

I think I understand why OP thinks the save process should work as described but (a) I don’t think it would work for Scrivener and (b) Scrivener already works as described after a fashion, just not exactly like “Windows norm” products.

One key misunderstanding is a manual “Backup Now” is by default given a different file name than the rolling auto-backups. (1) So Scrivener has Save and auto-save saving to the same location, the project directory the user is working from; this is like Word, etc’s use of temporary files that the user works in (2) The user can create a manual backup with “Backup Now”; this is like Word, etc’s use of manual saves. (3) Rolling, configurable auto-backups makes sure the work at close out time (or whenever the user has it configured) is always saved; Word, etc tend to not have this option and is very helpful for either the "whoops, I meant to backup/save before I exited times or just always making sure a backup is made (such as I use, to Dropbox, so it is saved offsite). I didn’t even mention Snapshots, which is again another function other applications don’t have (aside from a Save As… function, and then Word can’t do Save As… for only a partial section of its document).

The mechanics of how Scrivener does it may not be the same as other Windows word processors, but my opinion is that it is both superior and still maintains the same functionality and, in fact, even further functionality with how configurable it is. I think the OP needs to learn about how Scrivener was programmed to fit his/her expected use. Yes, with auto-save the “Save” menu item is essentially worthless (actually, very useful if the user configures auto-save to longer), but the Backup Now and rolling auto-backups will allow the user to save and backup work at any interval needed.

There is one thing that Scrivener is harder with, as another poster pointed out. If the user doesn’t like some recent change, in Word, etc. they can hit X and reopen to the previously saved point. In Scrivener they’d have to X out and find the file they last used for “Backup Now” (or look through the rolling backups), copy that backup to a working location (or, really, to be like Word, exactly in place of the existing directory) and open from that. As mentioned before, this behavior and ability to quickly X out and reopen is only more advantageous if you are more often making drastic, undoable mistakes that you need to revert to. I actually do need that in my secular work where I often open and edit documents that I don’t want to save any changes to.

Besides that, the OP just doesn’t like the terminology Scrivener uses.

To the OP, I suggest that you use Scrivener as it is. You’ll get used to it. After all, you got used to Word.

You might try one of the Open Office suites, libreoffice.org/ or
www-03.ibm.com/software/lotus/sy … e.nsf/home

Note, as others have said, that Scrivener has requirements beyond the single file when saving. Depending on the change, such as an option change, a title change, a drag and drop in the binder,
a link added/removed or the addition of a resource, the project file and other files in the .scriv hierarchy change, not just text files with your property within. To make Scrivener work ‘like Word’
would mean essentually doing a Save As at each opening of Scrivener and then, at the close, either reintegrating the changes into the original hierarchy or switching to the saved hierarchy and removing the original. This would result in 2 long periods of delay and, I my opinion, 2 points of high risk. I do not want to see that in Scrivener.

Someone up there said the Scrivener team had ‘nailed’ save and backup. Maybe so. At least they’ve made an excellent first effort. But something as important as save and backup deserves to be ‘glued and screwed’. So here are a few suggestions for improving backup.

Save works well; I don’t see a need for changes. I like the little * in the title bar indicating a file is ‘dirty’. It goes away, after the idle delay, probably 2 seconds, to tell you all is well, your property is safely on disk. Note that things like option changes don’t show the *. They appear to be saved immediately; a good thing.

Backup has some flaws.
a)
I find the ‘keep so many’ scheme inadequate. It doesn’t leave enough historical backups.
I’d like to see it changed to have additional control (and supporting logic, of course):

  • Save the last so many daily backups.
    This would mean that Scrivener does not delete the last backup done yesterday.
    The limit would be set at 30 by default.
    This means you’d have 30 days worth of backups, not only the 25 you did just yesterday.
  • Save the last so many monthy backups.
    This would mean that Scrivener does not delete the last backup done last month.
    The limit would be set at 12 by default.
    This means you’d have 12 months worth of backups, not only the 25 you did just yesterday.
    You’d also have the daily backups and the backups done today.
  • Save the last so many backups.
    This remains to limit the backups done today.

b)
When I started with Scrivener I, wrongly, assumed that backups were project related. By that I mean I thought they were saved somewhere related to the project. But instead I found that all backups are written to the same directory. This is documented but I only noticed it because I went looking for it. There’s also nothing on the GUI to indicate that the setting is ‘global’ to all projects.

The problem is that they’re all together in there and if you’ve somehow copied a project somewhere and it has the original project name then the backups are intermingled. And the limits apply to all projects of the same name. Which means you can lose all your backups for some projects just by working on a different project. Clearly some ‘glue and screws’ are needed here:

  • Save backups based on the project path, not just the project name.
    An alternative to this is to make the Options/Backup/Backup location option “non-Global”.
  • Backup limits apply on a project path basis.
  • Keep a flag somewhere in the project indicating the timestamp of the last change.
    Use the flag to skip backing up when there’s been no change.
    This is only for automatic backups.

c)
I’ve experienced several disk crashes. Because of that I now work almost exclusively on external hard disks. And I use a laptop. Even though I backup everything once a day I’d like to see:

  • Option for 2 different Backup locations.
    Ensure they’re on different disks/devices.

I have yet to try the snapshot feature so I can’t speak to that.

Al