I think it’s fascinating to see all the different ways that techfolk who write fiction set up their writing pipelines. Some people go with a bare-bones coding editor like Vim; others use “pro” tools Scrivener. And then there’s the typesetting pipeline. Using troff is pretty out there.
When I started writing fiction, here’s what I came up with. I write in Google Docs because it’s lightweight, works everywhere, and can easily export to Markdown format. From there, I manage the Markdown source in Git and Emacs. For typesetting—I cannot abide crappy word-processor output—I used to use Pandoc to generate LaTeX markup and run LuaLaTeX to produce PDFs. But that took several seconds. Now, I use Typst to go straight from Markdown to PDF files. It’s pretty much instant.
If you want to see what the output looks like, here’s the first chapter of my current project:
(Notice: This is a typesetting sample, not a writing sample. It’s my W40k-adjacent fanfic, so you have been warned.)
It’s set in 12pt Erewhon, which is an enhanced version of Andrey V. Panov’s Heuristica font family, which extends Robert Slimbach’s Utopia font family. Slimbach is a sublime type designer—he created Minion—and Utopia is one of his early works that is especially spirited.
Huh. I wrote into vim and made a bunch of .md files, and then just had Claude assemble them into an epub to make my book for Kindle. How did you decide on that size of page for the PDF?
One writer recently wrote his own program for the job, but at least he eventually finished some books: https://hughhowey.com/neo/
Me, I just write in Emacs, set up a makefile, and export to PDF with Pandoc. But I admire the work to make a comfortable yet undistracting writing environment.
Indeed. One irresistible temptation for coders who go into writing is to bikeshed their pipeline. I know I have spent many hours tweaking my writing pipeline when I should have been writing. I mean, why just write when you can create a Containerfile to build a hermetically sealed image that houses your typesetting system and a dozen carefully chosen fonts, and then you can write scripts managed by a Makefile to launch containers from the image and inject into them your source text and extract from them the typeset PDFs? I mean, surely all of that is essential work on the critical path to getting your writing done. Right?
troff is no longer fashionable, and it's not what you want for top-quality typesetting, but I don't think of it as "out there". My non-technical mother was using an HDOS program called TEXT.ABS, which is similar to a limited nroff, to format the articles she wrote in the 01980s. It wasn't as easy to use as something like MacWrite, but it was a lot easier than a typewriter. The text editor she used was called PIE.ABS, and from my point of view, the usability distance from a typewriter to PIE was bigger than from PIE to MacWrite.
She didn't have much choice of fonts on the pre-Graftrax Epson MX-80 she used. I think it was Pica or Elite.
Typesetting is certainly very satisfying, but it's not essential to writing.
Agreed, it seems writers that also code have all kinds of different workflows. I end up using (x)html and doing print stylesheets and rendering pdf for print using weasyprint (works really well, and they have sample stylesheets on their website that are a good starting point).
I still use a custom markup language I made for myself [0] because I don't find Markdown flexible enough, Latex's syntax is disgusting, and neither is particularly consistent.
I typeset some university work in troff in 1994. The professor I submitted it to was impressed but questioned my sanity in front of half of the class. Not wrong to do so.
Might still have the source code to that work somewhere. And it probably still parses.
Can't remember why I did it but on reflection it might simply have been that LaTeX was not going to fit on my extremely space-constrained Linux machine but troff was already there.
ETA: this sentence made me chuckle. "It's much easier to read without the inline comments:" Oh yes, much ;-)
Might be worth mentioning that Tim O'Reilly himself co-wrote a book about typesetting with troff. Unix Text Processing by Dale Dougherty and Tim O'Reilly, Hayden Books, 1987. This text was subsequently provided free under the Open Books Project.
Larry Kollar has provided a recreated source code repository at
For those that don't know, groff's ancestor "roff" is very closely tied to UNIX's origin story. roff itself had it's own antecedent called RUNOFF (whence the name) on Multics. roff was basically UNIX's first "killer app" and was used to typeset patent applications by the division of Bell Labs that handled that sort of thing. It legitimized the further development of UNIX.
Groff seems to be somewhat supplanted by mandoc these days, at least on the BSDs.
I was glad to see this page about groff, which is much underrated as a publishing system. The perl script for replacing some markdown markup with chosen groff macros is very interesting. I was however wondering about why the OA chose to write their own macro set rather than simply using the mom macros? Was it simply the need for German markup (and I take the point that languages other than English is an issue)?
Nils Holm (https://www.t3x.org/) also publishes ebooks using groff, but he does not (so far as I am aware) publish his markup or macros.
As to the writing: I'm still working on the sentences, so 45 pages is a distant goal.
One thing I have noticed about S-tier writers is that most of them have an established routine and stick to it. Why risk breaking something that works? With an established routine, you settle into it, and muscle memory puts you into writing mode. Start mucking with the routine, and that is how you get writer’s block.
Quick search shows George R R Martin wrote 3 to 4 million published words. ASOIAF is ~1.8 million. That's enough to focus on in my opinion.
Honestly, the amount of writing doesn't even seem that useful of a metric. Stephen king only wrote ~3x more than George, and about 6x more than someone like Nabakov (1.5 to 2 million words across two languages). The real takeaway IMO is to just write regardless of the technology you use
While he only did it, to my knowledge, for one book: Stephen King is the only major writer I have heard of to write a full novel with pen and paper (the novel in question was _Dreamcatcher_), which is really interesting to me.
A quick search immediately expands your knowledge (and unfortunately makes Stephen King much less interesting in this regard): Steinbeck went through 300 pencils to write East of Eden. Nabakov wrote all of his novels entirely on index cards. Even JK Rowling apparently wrote the vast majority of harry potter by hand on unlined paper.
And while not a novelist, Tarantino famously writes his screenplays all by hand as well.
Sorry to break it to you, but there are many, many novelists who prefer to write longhand!
Some notable examples I remember include J. R. R. Tolkien, Patrick O'Brian (who wrote all 21 books in the Aubrey-Maturin series by hand, and claimed at one point to write with "ink and quill"), Nabokov, Joyce Carol Oates, Ursula LeGuin, and Truman Capote.
I think 9front (Plan 9 fork) dev, cartoonist and fiction writer Stanley Lieber also formats a good part of his stuff with troff, e.g: https://massivefictions.com/2024/06/24/0/
"This book was typeset (troff -ms|lp -dstdout|ps2pdf) in URW Garamond No.
8 by the author, using an MNT Reform laptop running the 9front operating system."
>Building a world with its own set of rules and history and what not turns out to be very fascinating. ... The length of the stories make proof-reading ridiculously hard.
Sometimes I wonder if writing more short stories would be like playing 9x9 Go games for new authors. It won't exercise the big structural challenges, and it comes with its own challenges, but it does practice a lot of this in a much more digestable size.
As far as world building goes, I think GRRM/Rothfuss/Sanderson style "world building through epic scales of structures and systems kept consistent" is overrated and maybe even Considered Harmful if you look at what it's done to my first two examples. By contrast, Borges does more world building with the ontological tricks in his <20 page short stories than it feels like Rothfuss ever did, in much the same way Tolkien did it with similar tricks. For example, treating himself as a mere translator and even doing some philologist tricks like telling you he's simplifying some names into English (Sam's real name? "Banazîr"), and withholding more information rather than dumping it on you, means that his stories feel like recovered translated works from a world just as real as ours and that's why people still devour The Silmarillion just to get more of a look into it.
Borges would do similar in his short stories, even embracing a little bit of inaccuracies in the (fictional) narrator's account of some events, then hitting you with a surprise endnote at the end of the story with scholia about the provenance of the story given that narrator's mistakes.
Wolfe was famous for it too. Embrace some of the challenge of getting the details right and if you frame it right, it makes it more human and more "real".
At my dayjob, our legacy customer communication generation (for letters and such) has a runoff syntax. It's z/os, so not groff, but something of that lineage. People are still working with those documents to this day, generating millions of them a year. They all look at me funny when I identify it as runoff, everybody else thinks it's something custom in-house.
Back in the day I was part of a 'commission' to produce a 5-yearbook for our student club. It being the digital age and we studying at the premier (and only...) agricultural university in the Netherlands I went and had a look at the possibilities to use university facilities to typeset the book. Those facilities ended up being a PDP-10 running TOPS-20 'cause the VAXen were too important for such work. After some experiments with TECO and RUNOFF I was convinced I'd never get the other members of the commission to work this way. Since my own C-64 was just that, my own machine and already in use for other things I ended up scrounging up an ancient tape-based word processing system consisting of a metal box with a tape drive connected through a thick cable to an IBM Selectric typewriter. While it wasn't the epitome of ergonomics it was more amenable to some of the more digitally impaired members of the commission. That the Selectric's motor burned out just when I was going to produce the master printouts which led me on a hunt for a motor (found one for ~$50) and a complicated repair (re-adjusting the whippletree linkage without documentation but with time pressure from the printers was not my favourite experience). Still, we made it, the book was published, all's well that ends well.
EPUB got specific structure/components, which includes metadata files, and isn't just `zip book.epub book.html book.css *.jpg`. Of course can set-up from start to have expected structure, write xhtml files as expected, write metadata files manually, and create epub with just `zip book.epub -r *` but is more convenient via script/app.
I think it’s fascinating to see all the different ways that techfolk who write fiction set up their writing pipelines. Some people go with a bare-bones coding editor like Vim; others use “pro” tools Scrivener. And then there’s the typesetting pipeline. Using troff is pretty out there.
When I started writing fiction, here’s what I came up with. I write in Google Docs because it’s lightweight, works everywhere, and can easily export to Markdown format. From there, I manage the Markdown source in Git and Emacs. For typesetting—I cannot abide crappy word-processor output—I used to use Pandoc to generate LaTeX markup and run LuaLaTeX to produce PDFs. But that took several seconds. Now, I use Typst to go straight from Markdown to PDF files. It’s pretty much instant.
If you want to see what the output looks like, here’s the first chapter of my current project:
https://drive.google.com/file/d/1Wl4k_rdYGFZp-r5nq864r5CNzD1...
(Notice: This is a typesetting sample, not a writing sample. It’s my W40k-adjacent fanfic, so you have been warned.)
It’s set in 12pt Erewhon, which is an enhanced version of Andrey V. Panov’s Heuristica font family, which extends Robert Slimbach’s Utopia font family. Slimbach is a sublime type designer—he created Minion—and Utopia is one of his early works that is especially spirited.
Huh. I wrote into vim and made a bunch of .md files, and then just had Claude assemble them into an epub to make my book for Kindle. How did you decide on that size of page for the PDF?
> techfolk who write fiction set up their writing pipelines
And perhaps infinitely procrastinating on the actual writing itself...
One writer recently wrote his own program for the job, but at least he eventually finished some books: https://hughhowey.com/neo/
Me, I just write in Emacs, set up a makefile, and export to PDF with Pandoc. But I admire the work to make a comfortable yet undistracting writing environment.
He just vibe coded it.
Well, that's kind of a good use of vibe coding. Let the clanker make the app and you focus on the art itself, the writing.
Indeed. One irresistible temptation for coders who go into writing is to bikeshed their pipeline. I know I have spent many hours tweaking my writing pipeline when I should have been writing. I mean, why just write when you can create a Containerfile to build a hermetically sealed image that houses your typesetting system and a dozen carefully chosen fonts, and then you can write scripts managed by a Makefile to launch containers from the image and inject into them your source text and extract from them the typeset PDFs? I mean, surely all of that is essential work on the critical path to getting your writing done. Right?
Not always:
https://t3x.org has two novels and tons of comp sci books probably made with very little...
troff is no longer fashionable, and it's not what you want for top-quality typesetting, but I don't think of it as "out there". My non-technical mother was using an HDOS program called TEXT.ABS, which is similar to a limited nroff, to format the articles she wrote in the 01980s. It wasn't as easy to use as something like MacWrite, but it was a lot easier than a typewriter. The text editor she used was called PIE.ABS, and from my point of view, the usability distance from a typewriter to PIE was bigger than from PIE to MacWrite.
She didn't have much choice of fonts on the pre-Graftrax Epson MX-80 she used. I think it was Pica or Elite.
Typesetting is certainly very satisfying, but it's not essential to writing.
Agreed, it seems writers that also code have all kinds of different workflows. I end up using (x)html and doing print stylesheets and rendering pdf for print using weasyprint (works really well, and they have sample stylesheets on their website that are a good starting point).
I like Erewhon. Nice font.
I still use a custom markup language I made for myself [0] because I don't find Markdown flexible enough, Latex's syntax is disgusting, and neither is particularly consistent.
[0] https://breuleux.github.io/quaint/
I typeset some university work in troff in 1994. The professor I submitted it to was impressed but questioned my sanity in front of half of the class. Not wrong to do so.
Might still have the source code to that work somewhere. And it probably still parses.
Can't remember why I did it but on reflection it might simply have been that LaTeX was not going to fit on my extremely space-constrained Linux machine but troff was already there.
ETA: this sentence made me chuckle. "It's much easier to read without the inline comments:" Oh yes, much ;-)
Markdown if you can get away with it, typst if you need it.
O'Reilly used to typeset their books using troff/groff: http://www.lemis.com/grog/Documentation/PUS/tmac.Gbignuts.G (“O'Reilly & Associates book macros by Steve Talbott, Lenny Muellner”)
Might be worth mentioning that Tim O'Reilly himself co-wrote a book about typesetting with troff. Unix Text Processing by Dale Dougherty and Tim O'Reilly, Hayden Books, 1987. This text was subsequently provided free under the Open Books Project.
Larry Kollar has provided a recreated source code repository at
https://github.com/larrykollar/Unix-Text-Processing
I recollect compiling the source from the github account having only to sort out one font path issue.
For those that don't know, groff's ancestor "roff" is very closely tied to UNIX's origin story. roff itself had it's own antecedent called RUNOFF (whence the name) on Multics. roff was basically UNIX's first "killer app" and was used to typeset patent applications by the division of Bell Labs that handled that sort of thing. It legitimized the further development of UNIX.
Groff seems to be somewhat supplanted by mandoc these days, at least on the BSDs.
The UNIX project was funded by management as a word processing system initially, yes.
I was glad to see this page about groff, which is much underrated as a publishing system. The perl script for replacing some markdown markup with chosen groff macros is very interesting. I was however wondering about why the OA chose to write their own macro set rather than simply using the mom macros? Was it simply the need for German markup (and I take the point that languages other than English is an issue)?
Nils Holm (https://www.t3x.org/) also publishes ebooks using groff, but he does not (so far as I am aware) publish his markup or macros.
As to the writing: I'm still working on the sentences, so 45 pages is a distant goal.
Game of Thrones author George RR Martin writes all his novels on an old MS-DOS machine, not connected to the internet, and he uses WordStar 4.0. [1]
[1] https://www.bbc.com/news/technology-27407502
One thing I have noticed about S-tier writers is that most of them have an established routine and stick to it. Why risk breaking something that works? With an established routine, you settle into it, and muscle memory puts you into writing mode. Start mucking with the routine, and that is how you get writer’s block.
You're right, but I feel compelled to point out the irony of this comment being re: GRRM, the world's most famous case of writer's block.
Well writer's block on one specific story. GRRM has written other stuff in the mean time.
Personally i feel for the guy, imagine the whole world insisting you only work on the project you started back in 1991.
Maybe should focus on someone like Stephen King or Brandon Sanderson who actually pump out writing.
Quick search shows George R R Martin wrote 3 to 4 million published words. ASOIAF is ~1.8 million. That's enough to focus on in my opinion.
Honestly, the amount of writing doesn't even seem that useful of a metric. Stephen king only wrote ~3x more than George, and about 6x more than someone like Nabakov (1.5 to 2 million words across two languages). The real takeaway IMO is to just write regardless of the technology you use
While he only did it, to my knowledge, for one book: Stephen King is the only major writer I have heard of to write a full novel with pen and paper (the novel in question was _Dreamcatcher_), which is really interesting to me.
A quick search immediately expands your knowledge (and unfortunately makes Stephen King much less interesting in this regard): Steinbeck went through 300 pencils to write East of Eden. Nabakov wrote all of his novels entirely on index cards. Even JK Rowling apparently wrote the vast majority of harry potter by hand on unlined paper.
And while not a novelist, Tarantino famously writes his screenplays all by hand as well.
Sorry to break it to you, but there are many, many novelists who prefer to write longhand!
Some notable examples I remember include J. R. R. Tolkien, Patrick O'Brian (who wrote all 21 books in the Aubrey-Maturin series by hand, and claimed at one point to write with "ink and quill"), Nabokov, Joyce Carol Oates, Ursula LeGuin, and Truman Capote.
But what does he use to not write them?
I think 9front (Plan 9 fork) dev, cartoonist and fiction writer Stanley Lieber also formats a good part of his stuff with troff, e.g: https://massivefictions.com/2024/06/24/0/
"This book was typeset (troff -ms|lp -dstdout|ps2pdf) in URW Garamond No. 8 by the author, using an MNT Reform laptop running the 9front operating system."
Well, sl might have patched troff itsef....
Whoah, wait, author uses groff without ms?! Even wrote his own macro package. Neat!
IIRC the Go book by Donovan and Kernighan was also written in groff.
>Building a world with its own set of rules and history and what not turns out to be very fascinating. ... The length of the stories make proof-reading ridiculously hard.
Sometimes I wonder if writing more short stories would be like playing 9x9 Go games for new authors. It won't exercise the big structural challenges, and it comes with its own challenges, but it does practice a lot of this in a much more digestable size.
As far as world building goes, I think GRRM/Rothfuss/Sanderson style "world building through epic scales of structures and systems kept consistent" is overrated and maybe even Considered Harmful if you look at what it's done to my first two examples. By contrast, Borges does more world building with the ontological tricks in his <20 page short stories than it feels like Rothfuss ever did, in much the same way Tolkien did it with similar tricks. For example, treating himself as a mere translator and even doing some philologist tricks like telling you he's simplifying some names into English (Sam's real name? "Banazîr"), and withholding more information rather than dumping it on you, means that his stories feel like recovered translated works from a world just as real as ours and that's why people still devour The Silmarillion just to get more of a look into it.
Borges would do similar in his short stories, even embracing a little bit of inaccuracies in the (fictional) narrator's account of some events, then hitting you with a surprise endnote at the end of the story with scholia about the provenance of the story given that narrator's mistakes.
Wolfe was famous for it too. Embrace some of the challenge of getting the details right and if you frame it right, it makes it more human and more "real".
At my dayjob, our legacy customer communication generation (for letters and such) has a runoff syntax. It's z/os, so not groff, but something of that lineage. People are still working with those documents to this day, generating millions of them a year. They all look at me funny when I identify it as runoff, everybody else thinks it's something custom in-house.
Looks like z/os has troff. Probably has a bunch of other BSD utility ports?
Yes, but troff/groff is not the only descendant of RUNOFF, and apparently they're using a different one.
Back in the day I was part of a 'commission' to produce a 5-yearbook for our student club. It being the digital age and we studying at the premier (and only...) agricultural university in the Netherlands I went and had a look at the possibilities to use university facilities to typeset the book. Those facilities ended up being a PDP-10 running TOPS-20 'cause the VAXen were too important for such work. After some experiments with TECO and RUNOFF I was convinced I'd never get the other members of the commission to work this way. Since my own C-64 was just that, my own machine and already in use for other things I ended up scrounging up an ancient tape-based word processing system consisting of a metal box with a tape drive connected through a thick cable to an IBM Selectric typewriter. While it wasn't the epitome of ergonomics it was more amenable to some of the more digitally impaired members of the commission. That the Selectric's motor burned out just when I was going to produce the master printouts which led me on a hunt for a motor (found one for ~$50) and a complicated repair (re-adjusting the whippletree linkage without documentation but with time pressure from the printers was not my favourite experience). Still, we made it, the book was published, all's well that ends well.
FWIW, to use ^H for bold print, set this:
or depending upon your shell. Useful for old printers like I have :)I applaud all human-based writing tools. 10 days for 40k words is very impressive!
I suspect many people will prefer HTML for markup. That's why I built EPublish, which is related but for an HTML -> EPub workflow:
https://frequal.com/epublish/
Isn't the HTML->ePub workflow called zip?
EPUB got specific structure/components, which includes metadata files, and isn't just `zip book.epub book.html book.css *.jpg`. Of course can set-up from start to have expected structure, write xhtml files as expected, write metadata files manually, and create epub with just `zip book.epub -r *` but is more convenient via script/app.
I think you can make the metadata files with roughly a five-line Perl script.
I have the impression that an HTML -> epub workflow mostly consists of making a zipfile?
Many people will prefer Markdown or Commonmark for markup.
>I suspect many people will prefer HTML for markup.
Most people hate writing HTML by hand. That's why we got lightweight markup languages like Markdown, Asciidoc, reST, etc.