Outline numbered lists behave inconstently

This thread went unanswered, so it seems worthwhile starting afresh since I’m using Scriv for Windows and older reports are different versions, OS etc.

Outline numbering (1, 1.1., 1.1.1., etc) seems incapable to keeping track of the indentation/tab settings when changing levels and I can only keep things consistent by periodically finding some good items and using format copy-paste. It seems there are many other manifestations of list problems, but this is the one I care about right now.

A couple of questions

  • Is this bug/are these bugs acknowledged?
  • Will the issues be fixed, and if so, when?
  • How can one set custom indentations etc.?

I really don’t see that fixing it should be hard because, well, outline numbering and indentation isn’t hard (I’ve implemented it myself a couple of times) and it’s a bit weird that the problem occurs at all and hasn’t been fixed already. Could someone explain what the problem is here?

1 Like

Lists in Scrivener are a known and acknowledged problem. They do not work well and can not survive even the most routine editing. Other than taking a “set it and forget it” approach to a list where you create it and never reorder the items, the best advice it to either avoid them, use markdown instead, or “treat them as unstable dynamite when editing around them.”

See these threads, among others:

Inconsistent formatting and indentation for hierarchical ordered / unordered / bullet lists

Reordering bulleted lists

@Mad_Girl_Disease I saw your “Inconsistent formatting…” post - that’s why I thought it worth starting a new thread. Hyperbolically*: the problem annoys us; on the off-chance that separately reporting the problem again is annoying enough to them to make them want to fix it, I’m in!

Frustratingly, outline numbering in MS Word is not without it’s challenges either, and I am trying to wean myself off Excel (one of the places I wrote my own outline numbering code) because: words & paragraphs, dammit - not formulae & cells.

* I’m really a nice guy and terribly supportive and appreciative of their work, but sometimes I feel that nobody loves me and I just want to lash out :slight_smile: Hmmm. A new piece for Medium? “Outline Numbering Trauma… 13 things you can do to de-stress: 1, 1.1, 1.2, 1.2.1…”

Update just read @AmberV on undo etc. and handling dynamite. The thing about dynamite is it’s the safe way of packaging nitroglycerine :rofl:

1 Like

I need more of a report than this to work from, I’m not sure what the expectation is, or how to trigger the conditions you refer to where consistency breaks.

How can one set custom indentations etc.?

Perhaps this is a clue, if you’re trying to get fancy things might get messed up. I would treat it more as you would most things in Scrivener: data input, not desktop publishing layout. Just get the information into the structure you need, and worry about formatting in post. Honestly I wouldn’t even bother with numbering types, but that’s probably a bias coming from my Markdown background. I don’t see the point in a writer worrying about even 1. 2. 3. vs a. b. c. Let the output template handle that menial labour. :slight_smile:

Even so, if there is a specific bug we can fix to improve the more risky way of using lists, it would help to know exactly how to see it.

I’ve reported this one a while ago:
Bug: Indented lists in e-books can’t jump back from level 3 to level 1, and the ordered list <ol> is not closed. Using different levels 3 in a list leads to a list that indents further and further. I’ve had one that was numbered: Added a Raw HTML style containing </ol></li></ol></li></ol> in between the text as a work around. Corrected this by hand in Sigil post compile.

I may have my wires crossed, as I was thinking the issue they were referring to was relating to the editor itself, how text is formatted, and potentially breaking in the editor, rather than how any of that exports to HTML elements, rightly or wrongly.

Happy to take a look at HTML export bugs too, but I couldn’t find your previous report. The only post by you including the words “list” and “html” are this one.

Thanks @AmberV

Hope this helps, I’ve made it reproducible. TL;DR: key info isn’t being saved… problems occur after quitting and restarting. Details follow…

Here are some steps to reproduce this specific bug

  • Create new text document
  • Click in it and choose list style “1 1.1 1.2 2 2.1”
    • Number 1 appears in document - but note that the list style indicator says “1. 2. 3.”
  • Type words, hit return. New para numbered 2 appears
  • Type words, hit return
  • Click in para 2 and Home key to beginning of line. Hit tab. Item 2 becomes 1.1, 3 becomes 2.
  • Go to end of 1.1, hit return. Para 1.2 created. Type words, hit return, 1.3 created.
  • Type some more words in 1.3
  • Click in 1.2, hit Home, then Tab. 1.2 becomes 1.1.1, 1.3 becomes 1.2

Current state is this


Lots more editing, binder navigation, type in other editor window… it won’t go wrong!

→ Quit Scrivener, restart…

Type text in 2.4.2, hit return ant tab and indent is now wrong - it was number starting just under the text of the higher level para above


Shift tab and it’s wronger


But if I start a new outline number list below the new heading the numbers there are fine… and still persist in wrongness in the older set


Quit, let it take backup, restart

After restart in the “latest” list

Tab here and it’s wrong again



Something isn’t being saved correctly. Fix that and who knows how many problems will go away! :smiley:

PS I expect there are other circumstances that could cause that key info to get lost… e.g. after sufficient time old data structures could be garbage collected and Scrivener will go to disk for the saved text?

PPS After restart, new paras at the same level are OK, it’s only changing levels that seems to go wrong.


Thanks! I can very easily reproduce the problem now and get it written up (it doesn’t look like it ever has been).

1 Like

It should have been, though. Was first noted in the beta process very early in and wasn’t really addressed then, so hasn’t been addressed at all. :frowning:

Same holds true for the tables issues noted in the beta. Some have been fixed (sort of). but people are still reporting the same issues from a year or two ago that haven’t been addressed.

If you’re saying that there are more people creating posts on the forum than one or two people can come across or fully address, I would agree.

Hello AmberV, Thank you for your input. I have struggled with trying to get consistent formatting of lists for some time. However my challenge has not been deciding between “1. 2. 3. vs a. b. c.” lists. It is the question of consistent indentation when trying to make a number list or bullet list with no indentation in front of the number of bullet, and minimal indentation afterwards. Or even trying to change the indentation at all. Is this what you are referring to when you said that you “wouldn’t even bother with numbering types?” It seems I have wasted untold hours over the years trying to force Scrivener to do something it isn’t made to do.
I like to use my phone to make presentations I have created in Scrivener and saving that little bit of extra space is helpful.
Not being too technically minded, It sounds like you are saying that Scrivener is not really intended to be used for that kind of thing. Am I correct?
Thank you for your help.

I don’t care much to discuss numeric vs alphabetic, but when 99% of the page is alphabetic, numeric provides a greater visual contrast, and it’s a lot easier to read that a. b. a. c would be.

What I really care about is that outline numbering, so to clarify the use-case to make the value of 1, 1.1, 1.2, 1.2.1 (suitably indented ) clear: I am currently bashing in plot development for the 2nd half of a thriller. At entry time items are largely sequential, but some points are definitely detail so they can be demoted, and some are major plot points, initiation, crisis… and need to be promoted if they are entered while I am down in a lower level.

Being able then to adjust the level with tab/shift-tab is indispensable and the use of documents, cards is no substitute for the convenience and co-location of all new items.

Once I have then reviewed and assigned levels the structure then leads itself to the creation of documents, sub-docs etc., but not before

And having it neatly indented is indispensable. Custom indentation also helps if one tends to use too many levels initially but still wants it to look neat and get as much text in as possible - currently (though maybe a side effect of poor indentation on resume) I think Scrivener indentation too generous given that I can change font size independently.


It seems that it’s not just the indent that is getting lost between Scrivener sessions but also paragraph spacing, so I would guess all paragraph formatting is disappearing right now.

Another oddity:


@jac: Or even trying to change the indentation at all. Is this what you are referring to when you said that you “wouldn’t even bother with numbering types?” It seems I have wasted untold hours over the years trying to force Scrivener to do something it isn’t made to do.

You could say that the degree with which I do not format my text while writing, and rather state what I intend the text to function as, that it all can fall under the umbrella of that point you quoted. To clarify, here is what a few simple lists look like for me:

1. First
2. Second
3. Third

* First
* Second
* Third

Now when I say my lists look like that, I mean that, literally. I type in "1. " and then the text, press return and type in "2. " and so on. It is impossible for my lists to break because they are as solid as the letters of the alphabet I’m using to write the words into them. It does not even matter if I number them correctly, because I am only saying a general thing with “3.”, it is not “third list item”, but “a list item”. I could number each line “1.” and not bother. There are only these two types of lists—and so I am not even saying “1.” as oppose to “a)”, I am saying something much more general than that.

Now if you open up the user manual and flip through pages of that, and look at the various lists, you’ll see what appears to be a lot more going on than what I just described. I am being very, extremely, picky about the design of a list. I have spent probably close to a cumulative week of time tweaking the spacing and indent settings, as well as bullet selections and other aspects of their design. There are so many different formatting variations and such going on, in fact, that were I do attempt to execute this directly, using Scrivener’s formatting tools in the editor, I would probably require close to a dozen different styles, just to handle the differences in line padding between different sequences of elements and within different contexts (such as whether a list is within an already flush left indented block like a hanging-paragraph). All of that detailed variation and design vanishes once you look at the source text—which again pretty much all looks like the above example.

But here’s the key thing: none of that work ever needs to be duplicated, because it is a stylesheet that takes my simplistic “1.” in the creative writing space, and turns it into a formatted list item in a PDF. If I ever change my mind about a particular aspect of the design, I need only change the stylesheet and every single list (of which there probably close to 500!) is updated flawlessly for the next revision.

I just did that in fact a couple of weeks ago. I wasn’t happy with how much space was in between sequential paragraphs within one list item, and have adjusted the padding to better accentuate them. If I were using Scrivener like a word processor and messing with indents and spacing in the writing environment, then such a revision would have probably meant weeks of going through thousands of documents and applying formatting. And that’s of course pretending for a moment that rich text even really allows you to format multiple paragraphs into a single list item anyway, as they don’t really, whereas for me it is a matter of typing in the following:

1. First paragraph.

   Second paragraph.

   * And we can even go on to add a nested bullet list.
   * By following the same rules at all indent levels.

   Or conclude with another paragraph.

2. Next list item.
3. Third list item.

That is the essence of what I mean, when I say, I don’t even bother with any of this while writing. All I care about is: is this list better stated with a general marker for each line, or does it require enumeration as a sequence of thoughts or steps?

To return to what I said in that post though, I fully realise this approach is not for everyone—I am not saying, as a response to bugs in the Windows list tool, that everyone should switch to writing in Markdown. But I do think that Scrivener on the whole is meant to be used more like Markdown is, than it is meant to be used like Word or LibreOffice, where you must be meticulous about your formatting in the writing environment itself. If we are to put writing methods on a spectrum, where Markdown and other plain-text typing methods are on one side, and desktop publishing oriented stuff like LibreOffice is on another, then Scrivener would be a bit nebulous to place anywhere in particular, as there are multiple ways of using it (including Markdown, obviously), but its design intent is closer to the plain-text ideal than the DTP/WYSIWYG ideal. So given that, it makes sense to me to lean into that, rather than fight it, or try to make it act more like a DTP. It’s really not a good tool for that, being on both platforms a tool written using stock text editing components that are both notably fragile when pressed beyond basic usage.

@Julian_M1: I don’t care much to discuss numeric vs alphabetic, but when 99% of the page is alphabetic, numeric provides a greater visual contrast, and it’s a lot easier to read that a. b. a. c would be.

Exactly, and to conclude all of the above, I would also note that I made that initial comment before realising this was a bug report about a pretty serious bug in the Windows version that effectively makes multi-level list editing extremely annoying.

None of the above was meant to be a comment on that, but rather my initial impression of there being too much complexity added to list formatting, and finding that complexity fragile at a later date—which sounds more like what you’re talking about as well—that spending an inordinate amount of time on list design can sometimes be wasted since it gets lost or changed through other actions, or the act of exporting. I would hope that it is understood that at this point we’re just talking about theories and not whether or not your bug report is valid or something!

I might not myself agree with you on whether using the bullet list builder is a suitable substitute for outlining directly in the binder—that doesn’t make a lot of sense to me personally, but I suppose I get where you are coming from. It is less mechanical, in that Outliner or Binder-based outlining requires more keystrokes to execute. That is something I have a number of ideas for improvement on, as I firmly believe it should be as seamless as typing in a bullet list, and feel that would go a long way toward resolving the friction some have toward using the outline design for all phases of a project, rather than purely the structural implementation (often as a matter of necessity). Incidentally we did in fact experiment with use Tab/Shift-Tab in these contexts for handling indent. It is nice—but it kind of fell flat when you consider the Outliner’s multi-column capabilities, where it can be used as much like a spreadsheet as an outliner, and tabbing is a very natural way of switching between editable cells. It lead the brain to competing expectations and confusion.


That’s pretty much the same way that I use outlines, sometimes numbered, sometimes just bulleted, and the idea that such use is in any way “fancy” is, well, kind of fancy itself. :slight_smile: The plain nonfancy fact is that it can not be done reliably in Scrivener. :frowning:

That doesn’t jibe at all with my experience with Word. But even if it did, were my meticulousness to lapse while outlining in Word, at least I’d be able to do simple undos without breaking the entire list or table – which is simply not true of Scrivener. Is that not the “dynamite”?

That said, yes, my expectations for text editing in Scrivener are more like markdown or a plain text editor, like maybe about 2/3 of the way between Gmail and the original Google docs. But even absent outlining, when a single Scrivener document starts getting long, and I have no need to split it into multiple docs and work in scrivening mode (which is Scrivener’s great idea), but I want to do sustained detailed editing, I run sync in Scrivener and work in Word, then sync it back when done. And that’s fine. Scrivener’s usefulness and uniqueness are found elsewhere.

1 Like

Well, now that I’ve set that up (didn’t know it existed, didn’t think to look - and I thought I had at least looked at every single menu item), it seems rather a nice way of working, he said understatedly. Thanks! I can see I shall be learning scrivener for years :slight_smile:

Extra info for @AmberV: Scrivener’s placement of numbering w.r.t. paragraph indentation also seems off (I defined myself a set of Outline Level styles OL1-4 to rein Scrivener in ).

Now I have synced to folder and opened RTF in Word I see the outline numbered paras look different despite the standardness of RTF: in Word the numbering aligns with the left indent (as I would expect), in Scrivener, the body text is aligned with the indent + hanging, and the number is aligned in a way I can’t quite fathom (if one also looks at smaller numbers)

In scrivener…

In Word (the numbering is too long for the left indent to hanging space so the text moves right on the 1st line to the 1st implicit tab)

where we note the para spec (read off from word) is Left 1.7", hanging 0.35" (is there a way to see/read a para spec in Scrivener?)

IM(H?)O Word is right :wink: but I guess there might be thoughtful reasons for doing it differently - in which case explanation, so that the user can understand how to exercise control, would be welcomed because, for me at least, disorder is deeply distracting and right now I can’t see how to e.g. use word where it is more capable (e.g. numbering) and reliably get acceptable presentation back in Scrivener, so I do not yet have a solution for the outline numbering issue.

All suggestions welcome

sorry, this just posted to wrong place, so moved it…

I just downloaded today the newest version of Scrivener for Microsoft Windows (version, released October 18, 2022). I was prompted to do this when I first opened Scrivener today and there was a very helpful list of changes included with that prompt. Reviewing the changes, I did not see anything about correcting these malfunctions with the bulleted/numbered lists.

I’ve been looking at this discussion for a while now, and I believe I saw somewhere in it a message from staff Literature&Latte that they do not plan to correct this malfunction. Is this correct? I don’t see it now, though I have admittedly skimmed the thread today and could have easily missed it.

If it is not to be corrected, can you please provide an explanation?

A potential solution might be to re-use the code from the final version of Scrivener 1 for Microsoft Windows, as there was no malfunction in those versions of the program.

I am not a programmer and know nothing about code, so I’m sure this is much easier to say than to do. However, the above discussion, as well as all the other threads about problems with errors, malfunctions and otherwise funky and wonky bulleted and numbered lists should show how important lists are for writers. Personally, they are a primary tool in my planning process, and I use lists in the same way described by other posters to this thread. Further, I hear my same frustrations, annoyance and growing anger, from all the contributors to this and other threads about the bulleted and numbered lists problems. I, and I’m certain all the others posting about this, am frankly shocking that Literature&Latte hasn’t corrected this yet.

Hi @Writer05 ,

I’m just a customer, and not part of the L&L team. Below are only my opinions.

I don’t recall seeing this, and I very much doubt L&L would adopt that posture. See this recent thread that contains a response from L&L support to a question similar to yours:

Note the “QT framework” mentioned in Ruth’s post. L&L did not write Windows Scrivener from the ground up. Instead, they leveraged the QT development framework to do a lot of the heavy lifting. The upside of this approach is that it enables a tiny company with 2-3 developers to produce such complex, sophisticated software. The downside is, if there is a flaw in the framework, they have to devise ways to work around it.

That is why these problems with lists are so difficult to resolve, because it’s not a matter of the devs tweaking their own code–it’s the development framework that has the problem.

Windows Scrivener v3 is built on a different code version of QT then Scriv v1, and this is probably what’s caused v3’s list problems. But no, L&L can’t just copy the v1 code to v3–it’s not their code that’s the problem, it’s the framework’s code.

I used to rely on lists (like you) for planning activities, making notes, etc., but list handling even in v1 wasn’t great, so I found other ways to accomplish those tasks back when I was using v1 and started weaning myself off lists. The only lists I use now in v3 are very simple, typically one level. If I feel really daring and the list is relatively short, I’ll sometimes go to two levels. :nerd_face: I use the Outliner for anything more sophisticated than that.

The thread I linked to above has some other workarounds that you might want to try out. Perhaps you’ll find something that helps you.



Thanks, Jim. This is all very helpful to know. Especially why such a conceptually simple solution as copying code from ver. 1 and using it to replace the current ver. 3 code is not viable. And for why a problem identified so long ago has not yet been resolved. I’ll take a look at that link and see if it has anything for me. I’ll also look more into the Outliner mode – I haven’t looked at it much and, from what I have seen of it, it doesn’t meet my needs. But, that may just be due to the minimal time I have spent with it.

The Outliner is just a view of the binder, but with the ability to see a document’s meta data fields.

To use this as a substitute for bulleted or numbered lists, you’d create a folder to contain your list, then create new documents inside the folder, one for each item on the list. Indent as necessary. Turn on Outliner numbering if you like. See the screenshot from the Tutorial project below as an example of what it would look like.

It’s a different way of looking at lists. It might seem like overkill to create a document just to be an entry in a list–and I would only suggest doing so for complex lists–but it provides advantages. For example, for each item on the list, the document can contain explanatory notes or data; you can refer to list items from elsewhere using document links; you can bookmark them; etc.