GitHub for Books?


When I started writing my first book in 2003, I’d been blogging for more than three years. I’d learned the value of a conversation with my readers. Most importantly, I’d absorbed the obvious truth that they knew more than I did. So, with the permission of my publisher, I posted chapter drafts on my blog. The result was a variety of comments and suggestions, some small and some major, that in the end helped us produce a much better book.

That process was an early stab at bringing the Internet’s widely collaborative potential to a process that had always been collaborative in its own way: authors working with editors. The notion of adding the audience to the process was, and remains, deeply appealing.

The tools of online collaboration are still relatively primitive, and often discouragingly awkward. But they’re improving, and I’m seeing glimmers of hope that in a few years we’ll have vastly more capable systems.

As Charlie Stross notes elsewhere in this book, Microsoft Word, ubiquitous today for authors and their editors, needs to be replaced. I rarely use it myself, but there are times when it’s the only way I can communicate with an editor. (I prefer to write in a plain text editor and then, if necessary, format in LibreOffice Writer; however, I find Writer even less stable than Word.)

Screenshot from Poetica editing softwareThe Track Changes feature in Word (and Writer) is, of course, a primary reason we all use it. Google Docs doesn’t offer this feature. It should. The closest thing I’ve found on the web for this kind of collaborative editing is Poetica, an early version of an editing tool that recreates much of the style – and I believe value – of traditional editing.

But we don’t do just text anymore. We “write” in mixed-media formats, incorporating charts, videos and more into our work, and e-book formats still aren’t supported as well as they should be. I’m still looking, for example, for a great EPUB-native editor. The open-source Sigil is a fine start, but also very much a work in progress.

Collaboration is going to get a lot more complex. The most famous Internet collaboration is the one almost everyone uses, at least as a reader: Wikipedia. Editing isn’t terribly difficult, though not nearly simple enough for true newbies. Even if it was, Wikipedia isn’t a book with an author’s voice, and isn’t meant to be. Yet it shows many of the ways forward, including the robust discussions in the background of the articles.

Wikipedia articles are also living documents, changing and evolving over time. Could books be like that?

They could in the editing process if we use powerful tools from the software world. I’m thinking here of GitHub, the version control system used by many software teams. What might a book look like created in GitHub? A team at Princeton’s Institute for Advanced Study has shown us with a dense (to non-mathematicians) volume called Homotopy Type Theory: Univalent Foundations of Mathematics. As noted in a story about the project, this was more than just an enjoyable project for some reasonably geeky folks: “If they’d tried to write this book by emailing each other files or using something like Dropbox, it would have been a complete mess…. But GitHub made it fun” (McMillan).

At least one writer (with programming skills) is working on a project to make this kind of collaboration easier than it is with GitHub. It’s called PenFlip, and described as “GitHub for Writers.” I’m signing up for the beta.

If books are to become living documents after their original publication – and I believe they should in many cases – we have another major hurdle: the book-numbering system called ISBN, or International Standard Book Number, a unique identifier created for commercial purposes. But the Library of Congress insists that any significant change to a book requires a new ISBN number – and that system is controlled by a single company that charges extortionate rates for individual authors.

There’s actually a good reason for this. If we cite a passage from a book, we need to know what version of the book we’re citing, not just what page (or URL if it’s posted online). Wikipedia archives every edit made to an article, and you can cite any version of the article you choose.

It will get complex, fast, to apply this notion to books. But in an era where some books can and should evolve, we should try. We should hack ISBNs, with or without the Library of Congress’ help (preferably with), and create a system that lets us constantly update our e-books and print-to-order physical books in a way that doesn’t break citations even as it gives readers the most current versions.