This is the transcript of a presentation I did at the Open Textbook Summit in Vancouver, on May 29, 2015.
How I got here: LibriVox
In 2005, almost a decade ago, I started a project that transformed how I thought about books. That project, LibriVox, inspired by Wikipedia and the open source software movement, asked volunteers to make audio recordings of public domain texts, and give those audio recordings away for free on the Internet.
LibriVox — famously — has very little in the way of quality control: the rules were, and have remained:
- the source text must be published and in the public domain
- the audio must match the source text
- the audio must be audible and understandable
And that’s it. If you have a stutter, we accept you as a reader of Twain. If you have a high-pitched girlish American accent, we accept you as a reader of Conrad’s Heart of Darkness. If you have a strong Indian, or Bulgarian or Chinese accent, we accept you as a reader of Austen.
In about 50% of the catalog, just one person reads the whole book — we call these solo projects. But in half of the catalog, 20 or more different people might contribute to reading a book.
In the same book, some readers might sound like BBC-trained voice actors; others might sound like nervous high school students reading in a second language.
LibriVox: All comers welcome
This is confounding to many listeners, but it’s the counterintuitive beauty of LibriVox, and its power:
We accept all comers, and consequently, and amazingly, instead of having a catalog filled with unlistenable audiobooks, we have half of our almost-9,000 books read by a soloist, some of whom are also voiceover professionals; perhaps 10-20% of the catalog is difficult to listen to. Most of it is good. Some of it is excellent.
And beyond this question of selecting narrators, we accept any (published) and that is in the public domain: we’ve got Thucydides, Kant, Bronte, and Einstein, as well as Darwin and Austen and Joyce and Matsuo Basho; we have works in Cantonese and Church Slavonic and Urdu and Old Sudanese. We also have an audio version of “The Romance of Rubber,” by the United States Rubber company, and the number “pi” read to I am not sure how many digits. In Welsh.
LibriVox always has been, and still is a crazy kind of project — though I guess these days, anything-goes cultural production is the norm. Still, there is something very interesting about LibriVox in how complex and collaborative the project is — with teams of people getting together to make audiobooks, to usher a work from “hey let’s make a recording of Ulysses” … to a finished work of some 29 hours of audio — projects that last, as our first recording of Ulysses did, sometimes more than two years from start to finish.
Of course, no traditional publisher would ever have considered undertaking a project like LibriVox for so many reasons: the lack of quality control, the lack of business model and the lack of belief that so many people would be willing to do for free the kinds of things publishers do for money.
And yet here we are almost 10 years later, and LibriVox still produces about 1,200 new audiobooks a year, has a catalog of almost 9,000 titles, many of them of professional quality – or near enough anyway. Some of them less so. Our most popular books have been downloaded more than a million times each.
The last time I checked with Brewster Kahle, of Internet Archive (where we host our audio), he told me LibriVox books had been downloaded 250 million times.
We have thousands of volunteers, and we get a million or so visitors to our site every month. We are not Wikipedia, nor even Gawker, but for a project with virtually no budget, we’ve done something, I believe, important in the world.
Inside LibriVox: The things that make it tick
There are a lot of things about LibriVox that are interesting to me:
- We built a completely distributed workflow for multiple teams, doing multiple things, to work on creation of audiobooks, all on free, open source software (which we eventually stitched together into more custom software)
- We built a self-organizing management system to allow teams to coalesce organically around a book, with a kind of volunteer commissioning editor/production editor — which we call a book coordinator; volunteer narrators, who take on the chapters they want to read; prooflisteners who verify that the text indeed matches the audio; and a “publisher,” or metacoordinator, responsible for getting the metadata right and the books published out into our channels — in our case our website, and the Internet Archive as our repository of book files.
- We built a publishing platform that connects directly with listeners, that gets visited by a million souls a month — and generated a huge amount of direct-to-consumer traffic without ever doing anything more than what we were going to do anyway: publish free audiobooks online.
- We made a simple catalog API that allowed multiple app developers, and other services to use our catalog in their apps and tools — which makes it easier for more people to access our books – from apps in iPhones and Android, as well as other websites.
Books and value
But the thing I want to focus on today is this: we built a platform that allowed thousands of volunteers to build something meaningful to them on top of a fundamental “cultural asset,” on works that they loved … on the book.
And this for me has always been the most interesting thing about LibriVox: that the greatest value we provide is in many ways not to the people who might get free audiobooks to listen to, but rather to the volunteers who read our books aloud.
We give them a platform and a means to engage with books they love with both private depth, and public recognition. This passionate engagement drives these people to spend hours and hours of volunteer time working on audio versions of books, not for money, but for love.
That the world gets a free library of audiobooks has always been a kind of wonderful fringe benefit, a side benefit beyond what we really do (as I see it anyway), which is: to help people make audio recordings of public domain books, and give them away for free.
Massive multiplayer literary online game (MMLOG)
In fact, I sometimes like to think of LibriVox not as a publisher of free audiobooks, exactly, but rather as something like a massive multiplayer online game, a kind of literary World of Audiobookcraft, where guilds form around a particular objective — transforming a given public domain text into free audio versions — and can level up through recording and editing a chapter, from prooflistening, from managing the process, creating a cover, uploading a book to the Internet Archive and publishing the metadata on our site, and into scores of Audiobook apps through our catalog API.
I’m here to talk about Open Textbooks of course, but I bring up LibriVox because it is my touchstone when I think about new kinds of models for publishing books — and the kind of things people can build using simple digital tools, and an accessible library of ebooks.
In our case we started with the public domain works on Project Gutenberg — one of the first sites on the Internet — which has a library of free ebooks … critically in a simple and manipulable format — plain text or simple HTML.
The process at LibriVox goes is omething like:
- Someone decides they want to create an audiobook of a particular book, say “Howard’s End” …
- This person will announce that we’re doing this book on the LibriVox forum, and provide a link to the book at Gutenberg at a particular URL …
- Joe volunteers to read chapter 7
- Jane volunteers to read chapters 8-10
- Joe and Jane go to Gutenberg.org, gets the text, copies their chapters into to a text editor
- and then they record their work, upload their files.
The key point is that the availability of a free, manipulable public domain ebook library is what made LibriVox possible along with some other underlying technologies.
Having easy to find texts, without copyright restrictions, in standard open formats on the web … is what enabled us to build a project whose purpose was transformation of text into audio.
Where Pressbooks comes from
A number of years ago, it occurred to me that while we had built this complex and wonderful platform for creating audiobooks, LibriVox — something similar ought to be done for creating books.
The thing that has always interested me is not technology per se, but the kinds of things that technology can permit, the ways in which technology can enable new models of creation, and collaboration.
When I started working on Pressbooks there weren’t any online platforms for creating books — that is, there was no easy tool that would allow a collection of people to work together online, to create a book, and get the files they needed for print (PDF) and ebooks (EPUB, MOBI).
There are of course other platforms doing similar things now … but this is what Pressbooks was built for — to make that technical part of creating books — the PDFs, the EPUBs, the MOBIs — easy.
But Pressbooks also has a kind of Trojan horse built into it: every book is, by default, native to the web.
And while regular authors and book publishers haven’t been much interested in this part of Pressbooks — I think it’s the most exciting part of the platform. It brings book content directly into the network, into the web — and so we can start thinking about networks of books in much more exciting ways.
I think of Pressbooks as a sort of base utility — part of the infrastructure that enables something much more interesting to happen (I think), which is creation.
The reason I wanted to build Pressbooks was to help the emergence of new kinds of book publishers.
An Open Textbook ecosystem
This introduction about LibriVox is more or less the talk I gave in San Francisco, at the Books in Browsers conference in October last year … and the second half of that talk was a lament about how constrained the ebook ecosystem has become, for various reasons. The traditional book publishing industry — not surprisingly — hasn’t been very quick to embrace new and radical ideas about connected books, and what we can start to do if we start to think of books as connected in the network.
As someone said after that San Fran talk, “Hugh, you sound so depressed!”
But this is why I am delighted to be here today at the Open Textbook Summit, because the group of you are here to figure out how to start building something so exciting: a vibrant open textbook ecosystem.
I cannot tell you what a pleasure it is to be addressing all of you — keen and excited and unconstrained — compared with the talk I just did on Wednesday at the big book industry event, Book Expo in New York, where I was lamenting again the inability of the traditional book industry to make space for innovation. And probably sounding depressed.
But today we are not here to talk about how hard things are, or how depressed I am about the state of the book industry; but rather we are here to celebrate and dream of a new future of open textbooks and OERs. A dream I think many of you are committed to helping become a reality.
Let me lay out some principles of what I think Open Textbooks need to be. I don’t think there is much controversial here, but here are 6 principles I think are necessary for the success of a truly vibrant Open Textbook Ecosystem:
- Open Textbooks should have a URI/URL
- Open textbooks should be available in HTML, on the web, and ideally in as many other formats as possible — plain text, Word, EPUB, MOBI, PDF, paper, and whatever other formats people wish them to be in
- Open Textbooks should be open (not just to the students in the institution creating the books — but to everyone, everywhere)
- Open Textbooks should be free (online at least very cheap — though there is an important question of sustainability of the ecosystem that needs to be worked out)
- Open Textbooks should be remixable
- Open Textbooks should have good, standardized metadata, that is easily accessible online
So I will finish with a vision, a hope, a request to all of you:
Imagine if each of you left this summit with a commitment to create one Open Textbook at your institution. Just one.
Imagine if each of you do this publicly, online, and invite your colleagues and students who might participate.
Imagine if you solicited your students as editors — to critique, help along, and add to the chapters that future students will be using.
Imagine each of these books being made in a loose network, where you know John from Simon Fraser is working on Intro to Microeconomics, Sylvie from the University of South Florida is working on Intro to Macroeconomics, and perhaps Grace and Paul at the University of Botswana working on a translation, along with their own textbook on Credit Risk.
I have seen with LibriVox that a relatively small number of dedicated people — who were once strangers — can build a complex and prolific publishing program based on passion.
And while the mission of LibriVox is, I think, noble and good: “To make all public domain books available in audio, for free on the Internet”…
Open Textbooks have the potential to be so much more important for the world, by reducing costs for access to education.
The amazing thing about Open Textbooks is that we have the opportunity to do them right, to take every advantage of what digital has to offer, and to lead the way for innovation in our notions of what books are and what they can be.
So, to all of you I say: go forth, make Open Textbooks. Start with one. And then another. And, if this community does this right, I have no doubt that when we meet again in a decade at the 2025 Open Textbook Summit there will be thousands and thousands of Open Textbooks in the world, and all of you, I hope, will have played your part in making that happen.