Why Nabokov would use Org-Mode if he were writing today

Edmund Jorgensen (he/him) - https://tomheon.com, ewj@inkwellandoften.com

Format: 10-min talk; Q&A: BigBlueButton conference room
Status: Q&A to be extracted from the room recordings

Talk

00:00.000 Introduction
00:45.760 Nabokov's process of writing novels
02:24.080 Three practical problems novelists face
04:46.560 Org mode for writing novels
08:55.600 Takeaways and next steps

Q&A

Description

I've written several novels in Emacs. One of them grew into a monster with a baker's dozen twisty, interconnected subplots.

When I started to revise that novel, I had to use an outline to keep all the subplots straight, but I found it nearly impossible to keep that external outline consistent with the prose.

Finally I landed on a workflow using org-mode to keep the outline and the prose together, which significantly reduced the burden of keeping the two consistent as I moved and modified sections. I also found a way to use tags and sparse views over them to enable quick read-throughs of subsets of the book for continuity checks (which I plan to demo).

Later--long after finishing the book--I realized this process was essentially the Emacs update to the writing process that Nabokov used: he wrote on index cards that served as both prose and outline, so that he could move them around (which he did incessantly).

There's something deeply beautiful about org-mode's refusal to treat structure and prose as different things in a piece of writing--something I think Nabokov would have appreciated, and something I definitely appreciate, because it saved my novel.

About the speaker:

I'm Edmund Jorgensen, a software engineer by day and a writer by night, using Emacs for both. When one of my novels threatened to collapse under the weight of its own subplots, org-mode's powerful blending of structure and prose rescued it. I'd like to show you how that worked, and how much of org-mode's power for writing comes from its similarity to Nabokov's famous index-card-based writing process.

Discussion

Notes

• It looks like the Zettelkasten slipbox for nabokov
• James Howell also like the idea using small slide to convey single idea to the reader. In emacs, we have narrow function. Yes! I use various narrow functions to present text with Emacs. (I use narrow a bunch when editing, it really helps focus on a chapter or scene)
  • The funny thing about narrow functions, I mean the first time I saw it in the manual, there is a warning to the new user.  That would be afraid of this kind of functionality, and you have to be careful, haha...
  • I saw that warning too and avoided narrow for a long time as a result!  But it's not really that bad...
    • Exactly, I use narrow a lot, you know, every time I'm working on any single type of writing or writing a code or writing a piece of manuscript. It's really helped me to narrow down my attention and to kind of release any other thoughts that is not directly connected to the current things I'm working on. And that really is an underestimated functionality for the Emacs.
• The most valuable thing that Org will bring to the writer is the structure, how we can navigate between different structures of thoughts.
• The idea is using tag to narrow down a single person's timeline in the whole context of stories. It's something very interesting.
• ewj.io/emacs
• 👏 I'll start writing my masterpiece tomorrow!
• I need to use tags more, org-sparse-tree is handy

Questions and answers

• Q: Does the index really matter here? I mean, his colleague is also using some A4 paper, and do you think that the index card is the most important thing here?
  • A:
    • portbablity win!
• Q:How do you export the second level headings (scenes in this example) without the heading itself, just the content? 
  • A:3 ways for this: ox-ignore (it was visually annoying), dumb awk script, pandoc filters in lua
    • I would say the org-transclusion works very well for this kind of demand.
• Q: Slightly offtopic: where can we see your novels?
  • A: there are on Amazon: two of them, and a book of short-stories
    • Links:https://www.amazon.com/World-Enough-Time-Edmund-Jorgensen/dp/0984749233
    • https://www.amazon.com/Other-Copenhagens-Stories-Edmund-Jorgensen-ebook/dp/B00O4OQCBE
• Q: Have you looked at the Denote Signature features. The hierarchical nature of luhman IDs and index cards work well with Denote Signatures
  • A:I haven't, but I will take a look!
  • https://protesilaos.com/emacs/denote#h:f9204f1f-fcee-49b1-8081-16a08a338099
  • The part that I like with signatures is they can be optional with your zettelkasten as another way to use it.
• Q: Do you have a workflow combining hand-written index cards and org mode?
  • A:
    • Maybe just take a picture and OCR for your small index cards, but at the end of the day you always have to go back to your main Org files.
      • Ooh, I have a workflow for using Google's OCR to grab the text from my sketches (esp. the ID) so that I can link to my sketches in Org with ID and completion - sachac
        • haha, nice to see different approach, I personally didn't do that because I still most of my work is on the computer so yeah in the future if i have lots of handwriting notes in my working I will reconsider Google solution
• Q:
  • A:

Transcript

[00:00:00.000] Introduction

Hello, fellow Emacs enthusiasts. My name is Edmund Jorgensen. I'm a software engineer by day, but by night I love to write novels, and I lean on Emacs heavily for both of these activities. Today, I would like to talk to you about how Emacs, specifically with Org mode, has helped me manage some of the practical difficulties of writing long-form prose, novels in my case, and I'd like to get at this by talking about how another, much more famous novelist managed some of those same difficulties in a way that makes me suspect he might well use Emacs and Org mode himself if he were still alive and writing today. This talk will probably be of the most interest to listeners who either already write long-form prose in Emacs or are considering doing so, but I think that anyone with an interest in literature or Emacs will find something to take away.

[00:00:45.760] Nabokov's process of writing novels

So let's get to it. Here's a picture of a man lying on a bed, writing something on an index card. If we didn't know any better, we might think that he was just jotting down a recipe for beef stew or something like that. But in fact, this is not just any old man. This is Vladimir Nabokov, one of the most celebrated novelists of the 20th century, and he's not jotting down a recipe for beef stew in this picture. He's actually hard at work here, composing a classic of English literature on an index card. That's how he wrote all his novels, in fact, on index cards. I don't mean that he just took notes on these cards or wrote outlines on them. He did both of those things as well, but he also wrote the actual prose of his novels, word by word, sentence by sentence, on index cards. Let's see what that looked like at scale. This box you see here, full of groups of bundled cards, is what a novel in progress looked like for Nabokov. If you squint, you can see that these cards were from the composition of Lolita, probably his most famous novel. So why did he write novels on index cards? It's not necessarily an obvious choice. Yes, sadly, Emacs wasn't available to him at the time, but most writers in his day, if they weren't using typewriters, which were available, were using notebooks or loose-leaf sheets or something like that. Not these tiny little index cards. But Nabokov loved index cards. He swore by them because they represented an elegant solution to three of the most pressing practical problems that every novelist faces.

[00:02:24.080] Three practical problems novelists face

Writing a good novel is artistically difficult, of course. You have to write something interesting with a good story, something that people want to read. But writing any novel at all, whether it's good or bad, is brutally, practically difficult. You're hacking something like 100,000 words into unified shape over a long period of time, months or years. There are organizational challenges inherent in that process, and each writer needs practical techniques to manage those challenges. The most basic challenge, of course, is that, unless you're trying to bring back the Homeric Bard tradition of reciting books from memory in firelit halls, you need to actually set down those 100,000 words on some medium. In Nabokov's case, index cards worked fine for this. A little cramped, maybe, but workable. Secondly, as you're writing, you're bound to think of little but important things about the story that you want to record. I'm not talking here about big thematic notes or research that can go in a separate document, but smaller, more contextual notes that belong right along the prose that they refer to. These might be reminders, like, "Remember to clean up this sentence," or questions for yourself to consider during rewrites, like, "Why does Shirley feel this way here?" Nabokov recorded these notes in the margins of his cards or on the backs. Paper, in general, is great for this kind of intertextual note-taking. That's not particular to index cards. But what Nabokov really loved about index cards was how they solved the novelist's third and most difficult practical problem, which is imposing some kind of structure on this mountain of words. To have any hope of wrangling a novel into being, you need some way to break it down into parts, chapters, scenes, snatches of dialogue. You need some kind of higher-level outline that you can read, navigate, and rearrange as you consider and reconsider your story. You need structure. Index cards gave Nabokov a really powerful way to impose this structure because they created small, independent chunks of prose that he could bundle together into groups, like we saw in the box. This let him navigate his novel in progress quickly. He could just flip through those bundles, bundle by bundle, instead of card by card. He could also impose on and modify the structure of his novel just by shuffling those bundles around. So that's why Nabokov loved index cards for writing novels.

[00:04:46.560] Org mode for writing novels

Now I'd love to talk about why I love Org mode so much for writing novels and how it helps me tackle those same challenges. The first practical challenge, recording your words on some medium, is pretty simple. Org mode is a part of Emacs, a text editor, so you can just type in your text. We're not going to spend any more time on that. For the second practical challenge, recording small intertextual notes, Org mode offers comments, like this one here. The comment, "maybe I need to say which store?", with the leading pound sign there. I think that comments are generally underappreciated outside of coding. When writing fiction, for example, I love that Org mode lets me keep these comments close to the prose they refer to. I can see right here that I'm talking about saying which store in this first line, "One day, Bob went to the store." I get to keep these things close to the prose they refer to without ever having to worry that they'll accidentally be exported to a reader. That's great. So let's talk about how Org Mode handles the third and most brutal challenge of all, which is structure. Here we've taken the same text and we've imposed some structure on it. Like index cards, this is where Org mode really shines. Org mode extends outline mode, which is built around the concept of header lines, with different levels denoted by different numbers of leading asterisks (*). Personally, I tend to use top line headers as chapters and second line headers as scenes. You can see that here, where chapter one says "Bob and Shirley meet." Here's a scene, "Bob goes to the store." And here below is chapter two, yet unwritten, where Bob goes to work. Pretty exciting. Since Org mode supports folding, I can read quickly through a summary of my novel at either the chapter or the scene level just by flipping through different levels of visibility, just like Nabokov could flip through different bundles of cards. So here's the chapter level. I can see at a chapter level, "Bob and Shirley meet", "Bob goes to work." And then I can get one level more specific and see the various scenes in the chapter at the second header level. And I can, if I want, I can go all the way back to the prose level. And just like Nabokov shuffling his index cards around, I can move scenes around as logical units. Let's say, for example, that we wanted to move Bob's thoughts about life, which are down here, up further. Well, I can grab "Bob thinks about life," and I can move it up or down as a logical unit. But Org mode offers some even more powerful tricks for structuring and navigating your novel, beyond what even index cards can do. For example, you can use tags on your scene headings. You can see these here. They're the prominent colon separated words on the header lines. In this case, I'm using bob and shirley. These tags can represent characters who appear in the scene, which is what I'm doing here, or locations in which the scenes occur, or plot lines that the scenes further, really anything that you want. And you can then use Org mode's sparse view features to query a set of tags and trim your novel down to a subset of related scenes. For example, let's say we want to filter down to only the scenes in which Shirley appears. This could allow us to read quickly through just a subset of the prose, the prose that referred to Shirley in some way. Maybe we want to do that to check continuity for her character, or make sure that her character develops along a compelling arc, or even just to get a sense of how much airtime she gets in the novel. Thanks for listening to this whirlwind exploration of some of the practical challenges of writing novels and other long-form prose, and how Org mode can help tackle them.

[00:08:55.600] Takeaways and next steps

I'd like to leave you with a couple takeaways and next steps for those who are interested. First, if you're writing a novel or other long-form prose, or even considering doing so, take a look at Org mode, especially if you're already familiar with Emacs. It won't solve the artistic problem of writing an interesting book for you, not even with a ChatGPT plugin, but it's a fantastic tool for managing some of the practical challenges that come with hacking 100,000 words into shape over the months or years that that process takes. Second, if you're interested in learning more about some of the advanced features of Org mode and how they can help in this process, I wrote a long blog post about my difficulties writing a novel with 13 interconnected subplots, and how Emacs and Org mode saved it from imploding. I'll put a link here below. [ewj.io/emacs] Thanks for listening, and Emacs on!

Captioner: bhavin192

Q&A transcript (unedited)

seconds. And I believe we are live. Hi Edmund, how are you doing? I'm doing well, thanks. Yourself? his webcam on but he will be able to answer questions that you ask inside of the Azure pad that I've shared again on IRC. By the way, we only have 1 question and we have about 40 minutes of question time, so feel free to add as many questions as you want and in the meantime, we'll get started on the first 1. Unless, Edmond, do you have anything to say after your presentation? is the index, sorry, does the index really matter here? I mean his colleague is also using some A4 paper and you think that the index card is the most important thing here? I mean, I think you can do anything with a larger piece of paper that you can do with a smaller piece of paper. But I actually encourage you to try this out. I did, not for research for this talk, but just when I read about Nabokov and his index cards to begin with, I kind of tried it out a little bit and wrote some shorter things on index cards and so on and there really is something about the size and the kind of ability to manipulate them. You really can bundle them and move them around easier and I think that that I think he enjoyed that. So sure I mean I think you can do anything with a4 paper that you could do with index cards but I think there's something about that form that lends itself to the especially to the reorganization maybe to the focus as well just because it's smaller but but definitely to the reorganization. now. So thank you, everyone, for answering my plea for more questions. Next question. How do you explore the second level headings, i.e. The scenes in this example, without the heading itself, just the content? Is that clear enough? sorry, 3 ways with this and landed on 1 that I like. Originally I used the OX package. There's an OX ignore thing in there where you can add an ignore tag to where you don't want the headings, but you do want the content exported. I found that a little bit annoying, just visually annoying, when I'm, again, My theme here is navigating 100,000 word documents effectively and having that extra visual noise was kind of a pain. So I ended up, first I just did like a dumb ox script as part of my publication kind of pipeline that removed headlines at the scene level. And then actually, because I ended up leaning so heavily on Pandoc, and Pandoc, for those of you who have not looked at recent versions of Pandoc, they've got a really fantastic way to use Lua at this point to write filters. So you can kind of take the AST of your document and run these very simple Lua filters over it. They used to be in Haskell, which I'm not smart enough to write Haskell is 1 of the things that I've discovered. I keep bouncing off of it, but I'm just smart enough to write Lua. And so I use a Lua filter now, which I'm happy to publish to anyone who's interested. That basically lets me say, you know, what level headings to get rid of the heading, but publish the content. And part of the reason that's been useful is that some of the other novels I'm working on for example have different levels of hierarchy where maybe there's a part and then you know at the top level and then chapter and then scene and it's now the third level instead of the second and it's much easier in the Lua to just be like remove the third level headings or the second level headings or whatever it is so that's been that's been helpful. slightly off topic, where can we see your novels? they're on Amazon, there's 2 of them and a book of short stories. I think the short stories and the second novel, which is called World Enough in Time, which is the 1 that kind of prompted this talk, are probably of more interest to this, to the Emacs focused group. The first one's like a philosophical murder mystery, but the World Enough in Time is a kind of Douglas Adams inspired sci-fi comedy about kind of hijinks on a relativistic speed space cruiser, which was a lot of fun to write. It has a lot of twisty subplots, which is where I developed that technique of being able to filter down to tags and see a reduced version of the novel, which was very handy when trying to juggle 13 subplots. So yeah, check it out. links available on the talk page afterwards. Right now I sadly have to host so I cannot look up the links but we'll make sure or if Yeah. In the meantime we'll move on to the next question. Have you looked at the Denote signature features? The hierarchical nature of Lumen's ideas and index cards works well with Denote signatures. So are you familiar with Denote first? should check out. boxes. We talked a little bit about it earlier today. We talked about Orgroam, we talked about Denote as well as a lighter alternative to Orgroam. And yeah, the organization with index cards feels like it's something that would highly benefit from linking and back links and any kind of UX functionality for relating pieces of information. So yeah, definitely look it up. I use org-roam for a lot of different stuff and I would love, I will definitely check out Denote as an alternative. familiar with what Signature is within Denote and it'd be great if the person who asked the question could perhaps provide more details so that Edmund could get a little more information when he returns to the document. But yeah, if you're using Org-ROM, you're already within the mindset that you need, and perhaps you'd gain a little bit extra stuff from using Dino's signature, I assume. We have 8 minutes. We're still good on time. Next question, do you have a workflow combining handwritten index cards and org mode? I do write by hand when I get, I don't know what a good term for it is, I'll call it like editorial paralysis or something when I find it very hard to move forward in something because I keep going back and tweaking. And I will handwrite stuff at that point and then type it in because it's so much harder to get stuck in editing mode when you have to move forward on the page. I don't use index cards. In the blog article that I link in my talk, the ewj.io slash emacs 1, I did try using handwritten or spreadsheet outlines at 1 point and found them very, very clumsy for novel writing just because I do so much, I mean, I do so much revision that moving things around meant that I had to keep 2 things in sync with each other, the pros and the outline. And that was what really led me to Org Mode as a way to keep the, again, I think part of the key for me is keeping the outline and the pros right next to each other in a way that they move around which is just really, I don't know, for me really really powerful. questions available on the pad, but I see that some people have joined us on BBB, so hi everyone. If you have any questions feel free to unmute yourself and ask them. Otherwise, we might go on a break. So I'm going to give you about 10 seconds to unmute yourself. Or if you just want to add more questions on the pad, that's also fine. And that'll give you about 30 seconds. Otherwise, we'll need to go on a break. And in the meantime, I'll thank you, Edmund, for your presentation, because it's always nice, you know, we The reason why we have 2 tracks, and we've been having 2 tracks for the last 2 or 3 editions of EmacsConf is because it's really nice to have those talks which are still related to Emacs and to far distance developments because we are obviously using packages. But it's really nice to see when we foray into other areas like writing or any kind of academia-based topics. So thank you, it's really nice. It brings different colors to the spectrum of what EmacsConf is and what ultimately Emacs is as well. Thank you. thanks to you and all the other organizers for putting this together. Appreciate it. go on a little break for 5 minutes because I don't see other questions being asked. So everyone we'll see you again in 5 minutes and thank you again, Edmund.

Questions or comments? Please e-mail ewj@inkwellandoften.com