RTFM: Exhaustive search of Scrivener manual, including all 202 incidences of the word ‘import.’ The ‘Sharing’ portion of the ‘Preferences’ section was the most encouraging, but no.
Search the Forum: This link indicates someone having luck with comments, and seemingly round-trip: viewtopic.php?f=2&t=65580&p=333406&hilit=docx+import#p333406
Not a lot of others, easily found anyway.
Use Case: Compile novel to docx, submit to editor, re-import to Scrivener (before or after all comments resolved). So far I lose the binder layout, and I can only hand copy/paste each scene back, (and this can be heartbreakingly error-prone.)
Experiments: Many. I tried the “output links back to the different docs” and I got a flat single text import, with none of the folder hierarchy. Further compile options, now I get tags in the doc. I have attached the original project, the exported docx, and the new projects one split, one not, with the flat files and tags. I feel like I am close, but missing the exact formula.
Ok, here’s my experience.
I found it easiest by far to just leave my submitted work in Word format while going though the back-and-forth of editor comment rounds. I like Nisus Writer for this process, but YMMV. Word itself would presumably work, I just don’t own it.
Once we’d beaten it into submission (pun intended), I used the following process:
- Back in Scrivener, I selected all the document files as submitted and created named snapshots (“Accepted by Infamous Editor 3/9/2018”). Thus, I had a fall-back if something in the re-import went wonky.
- Next, I used the File->Import->Import and Split command to get the final Word version back into Scrivener, with the “Split using the document’s outline structure” option turned on. This got me almost all the way back to my original structure, with each section in a separate file inside a new folder in the Binder. A couple of sections had been merged or moved during the edit, but nothing major. (N.B.: Do make sure that in Scrivener->Preferences, Sharing panel, Import pane, that comments are set the way you prefer. )
- At this point, yes, I copy-and-pasted from each imported doc back to the original doc. A pain, but only done once instead of after each editing round. Because of import and split, matching up the import to the original was much easier than copy and paste direct from the Word doc. Now I have my final text in Scrivener, ready to, say, copy as a quote into a later installment of my story.
The disadvantage of this is that Scrivener just will not preserve those comment threads. Scrivener can’t deal with comment threads, so (as I recall) only the first comment in a thread is imported. I found I could either copy and paste each unimported comment (yuck) or create a PDF from the Word doc with comments in, and import that into Research. With this, I have the editing history preserved.
Hope this helps!
Thank you so much Silverdragon. Cool site, and stuff you have there! Nisus hunh? Takes me back to tower-shaped Macs! Cool that they are still at it. That workflow you described is pretty much mine, except for the import/split and then copying inside Scrivener.
is what gave me hope, but I can’t make it work. What I can do is parse the docx files with the various link options checked to try and understand what they add to the docx compile. That will be my next move.
In the files I attached before, I see the style tags, but they show up in the document. Okay maybe I need to import into a project that already has the styles defined. Ooooo, using the identical project I exported from worked! Getting closer. Now in this screenshot, (attached ImportNoSplit.png), you can see that the styles were applied! And, there are some sort of Binder directives ( [Open “Brunch”] ), etc.
Excited, I tried import/split, but the result is not great (attached as ImportYesSplit.png). The originating project is “TestIO.scriv” attached above. Compiler settings are shown in CompilerSettings.png (attached) The InsertUnique Unchecked.png (attached) is how the import looks if “Insert unique document identifiers only” is unchecked. (Oops max 3 attachments, all binder 'directives become the same: [Open in Scrivener]
Now I can see that the results of rtfd and docx are very similar. Both retain styling, though there is no connection to the Scrivener styles from the originals in RTFD, but the DOCX circuit properly brings each of my four styles back. So DOCX is definitely showing the most promise right now.
When I unzip and diff DOCX files with and without the link boxes checked, (see Screen Shot 2021-05-09 at 8.54.33 AM.png, attached) I can see that the DOCX with links checked has 21 more lines in it, so the above poster is correct about the intent of being able to round-trip with DOCX. That means I am not chasing a red herring.
I note that by default, (with nothing selected in binder) that the imports appear as one or more articles in Research. Perhaps this is the intent, but then what about those “[Open “Hello, I am a cookbook”]” directives? I imported into the draft, still got same result. Deleted all existing content, still got same result, one flat article with above directives in it.
Not giving up yet.
Oh, and just to be complete, I imported the turned off link options file, and found styles, everything, the same. Just none of those “[Open “Hello, I am a cookbook”]” directives.