UPDATE in Fairness to the Developers of Ulysses and others / I've Tried, I Really Have, But I Surrender 😔

[webwalrus]

Version control allows you to revert to any version from the past. So you could, for example, realize that you made some major mistakes with deleting content yesterday, and roll back to the version from three days ago.

The downside is that you have to learn how to use a version control system. This includes manual processes for adding, updating, retrieving, etc.

If you do not have the problem of frequently wanting to revert to a previous version, knowing what I know about you from the forums, this is just adding another layer of unnecessary complexity to your workflow.

And even if you do have the problem of wanting to revert to a previous version, if you can do that with your existing backups, do that.

[krocnyc]

@LisaSpangenberg

Styles are almost magical.

Agreed! I only wish that getting them set up were less fiddly so that more people would be open to using them regularly and rigorously rather than random inline styling.

[krocnyc]

@LisaSpangenberg

I abandoned Markdown because it felt like more work, not less.

I like Markdown for simpler documents that require some organizational structure, a minimal amount of formatting, and are amenable to being exported to another format via Marked 2. I can’t imagine putting together a complex document in Markdown alone—it wasn’t built for that.

[karlnyhus]

@Bmosbacker

I’ve only perused GitHub, but it has always struck me as very geeky.

I wouldn’t recommend a full blown version control system to you, for reasons you have mentioned. But, for heavens sake, use an editor that takes advantage of the Document Model that Apple introduced way back in macOS Lion!

In such editors, and you might already use one or more, the File menu has a Revert To command that brings up a Time-Machine-like interface that allows you to step back to older versions of your current document. This is one of the many reasons that plain text documents shine!

EDIT TO ADD an article from 2012 discussing the Document Model in macOS Lion and how it was improved in macOS Mountain Lion. (You may remember it as OS X 10.8 Mountain Lion )

The Very Model of a Modern Mountain Lion Document - TidBITS

[Bmosbacker]

Indeed, I’ve used this in Pages when needed. Scrivener has a similar but less intuitive and more manual process called snapshots. You can take as many snapshots as you want and revert back to those by document or entire project, but as far as I know, this is a manual, not automated process.

[karlnyhus]

@Bmosbacker

Indeed, I’ve used this in Pages when needed.

Well, then use that instead of fooling around with GitHub! Auto Save and Revert To were meant for people like you, power user that you are, and not GitHub.

[Bmosbacker]

I’m a “power user” in lowercase. If I don’t get this writing app thing resolved, I’m just going to be a “user.”

[Bmosbacker]

Scrivener has a similar but less intuitive and more manual process called snapshots. You can take as many snapshots as you want and revert back to those by document or entire project, but as far as I know, this is a manual, not automated process.

@karlnyhus I’m wrong. First time this year.

Scrivener can take automatic snapshots of changed text.

[SebMacV]

The author of this book produces a fun podcast on behalf of Literature and Latte. Most episodes lean heavily into the writer and their books, and talk about process. They end on some usage notes of how Scrivener has helped the authors write their books, and what are their favourite features. Most writers find something different they adore in the app. It’s a fairly low-key way to find out more about some of the app’s cool features, and if you are like me, hearing about the working habits of other writers and creatives keeps the writer in you going back for more.

[Bmosbacker]

Thanks! I just followed the podcast. I look forward to listening to it while I’m on the treadmill.

[SebMacV]

I am a ‘scholar’. Well, I aspire to the role, at least.

I write most of my short form notes in markdown (Obsidian, DT, sometimes Typora) and long-form academic work in Scrivener (RTF base format, I think). I edit my manuscripts in Word in engagements with publishers. It’s fine!

I hardly dare say it in such august company and among such computational experts with such finely tuned instincts about text production and editing on screens … but to me, it’s mostly just text. My markup in Obsidian is very light. What I need in Scrivener I just copy across (if you copy from ‘reading view’ in Obsidian, most formatting like italics or bold are preserved in copy/paste). The little bit of friction helps me keep the note-taking separate from the longer form drafting. I just find there are no real barriers between the two apps, and sometime wonder if we overcomplicate what is mostly just text. Most of the apps get the job of writing done admirably! Admittedly, some apps are better than others at organising longer form text (and so you can pry Scrivener from my cold dead fingers for anything over a thousand words).

[Bmosbacker]

That is a very good reminder. I could easily do the same thing using iA Writer or Drafts for that matter.
Thanks.

[rms]

@Bmosbacker

Scrivener has a similar but less intuitive and more manual process called snapshots

Better to use automated backups to zip files configured in Settings. On open, close, or set to make on each save. I auto backup on close, keep 25 copies and do not save on each save.

Simpler than manual snapshots and “save as”. KISS again.

[Bmosbacker]

@rms

Better to use automated backups to zip files configured in Settings.

Thanks, I already have this setup. I suppose snapshots except for an individual document are redundant then?

[rms]

@Bmosbacker

snapshots except for an individual document are redundant then?

That is what i think. I put my time into writing and not worrying about versions. In more than a decade of using Scrivener I never took a snapshot and if i had to roll back i took it from the relevant zipped backup. KISS.

With due respect you are worrying about the edges. Focus on getting the book written.

[Bmosbacker]

Yes, mother! In my defense, I have a daily target for writing, and am fairly consistent.

@rms I want to make sure that you understand that I really was trying to be funny. Your comment is a good one and I do take it seriously. I just didn’t want you to misunderstand my effort and humor. You have always been extremely helpful and I don’t want my attempt at humor to be misinterpreted. I would never intentionally be disrespectful or dismissive.

[LisaSpangenberg]

Snapshots are useful if you are about to make a major change to a ms., say reordering sections, or deleting a large chunk.

[karlnyhus]

@SebMacV

I hardly dare say it in such august company and among such computational experts with such finely tuned instincts about text production and editing on screens … but to me, it’s mostly just text.

+1

I especially look for apps that have commands like Paste As Plain Text and Paste and Match Style for the flexibility they allow.

[brookter]

May I make just one small amendment?

Snapshots don’t preserve the order of the elements in the binder, only the text within each section. You do need to back up the whole project before changing it if you want to be able to go back to the same order (and the same metadata state).

---

(Apologies if you already know the technical details, but I find that knowing what a project looks like under the hood sometimes helps understand why Scrivener acts the way it does, and I thought someone may find it useful…)

The reason for this behaviour is that under the hood the content of each folder and file in the Binder is in a separate Finder folder (with an impenetrable name) and the snapshots for each document are stored in a separate Snapshots subfolders with the same long name.

The order of files and folders is stored in the Project name.scrivx file (which is basically the project’s index), and that isn’t backed up in the snapshot process.

On the Mac we see the main project folder as a single Scrivener project, but it’s actually just a normal folder that MacOS treats specially (a ‘package’). You can see this if you right-click on a project in the Finder and Show Package Contents:

The content of each document in the Binder is held in the Files > Data subfolders with incomprehensibly long names… Any snapshots are in the Snapshots folder.

This is all held together in the Paper (APA) test.scrivx test file, which is basically an XML index file representing the current order of the Binder, as well as each document’s metadata.

So, to know which of the Files subfolders with the long names corresponds to any particular Binder item, you have to search for the title in the scrivx file to find the <Title> line, then look at the <BinderItem UUID="very long number" line above it.

So the only way of accessing a previous order (Binder state) is by restoring an older version, which isn’t done in the snapshot proces – that’s why you have to back up the whole project before making serious changes to the Binder.

It’s a bit more complicated that this of course (and I’m not an expert by any means), but that’s the general idea.

It can be interesting to look through the .scrivx file in a text editor, to see how things are held together, but please take a backup of the project first or you risk corrupting your whole project.

(Incidentally, Scrivener’s internal project structure is why Dropbox is the only Cloud app available for syncing Mac/Windows and iOS – none of the others apparently have the hooks to keep everything reliably in sync at the iOS end.)

BTW, this is all hidden on the Mac, but on Windows, which doesn’t use the same package mechanism, you just see the normal folder, with all the files visible. This causes problems when people try to copy the project, because they wrongly think that the Project name.scrivx is the whole project and they copy only that and then lose all their data…

Again, sorry for hijacking the post when you probably already know this, but I thought the details may be interesting to some Scrivener users.

[sangadi]

Backups are good only to the extent that can be recoverable. If there is data corruption then your backups are of no use and you will be surprised when you try to restore. Always restore and test your backups.

Also it’s not a point in time recovery. Though you may find some backups offer this feature but still it’s not a trusted system.

With a version control system it’s always a trusted system. Granted that the workflow will be more technical and needs some research in adopting a process. But you can recover from any kind of data loss, corruption and go back to any time in the past.

Enterprises rely on version control first and then add backups on top of that.

Granted that for this use case it might be an overkill but I still think that even implemented in a much simpler way will be more reliable, trustworthy and easily recoverable, portable, consistent than a traditional backup.