Friday, January 17, 2014

EPUB has a Steep Road Ahead (Notes from Open Book Hack 2014)

People I talk to about ebook technology belong to one of two camps.
  1. EPUB3 is the future of ebooks. 
  2. EPUB3 has lost the ebook format war because no one is supporting it. 

Here I'm helping Fortitude fix a bug.
(photo by Ray Schwartz)
So it was very enlightening to join a group of developers at New York Public Library last weekend at Open Book Hack 2014. Sponsored by NYPL Labs and the Readium Foundation, the event was convened to examine the challenges of the "Open Book" on the Web. Since I had written that another publishing hackathon "pretty much ignored ebooks", I felt compelled to attend, and write a report.

Open Book Hack was the first event I've been to where developers were actually working on EPUB, the format that is emerging as the underpinning of the digital book industry. This was both encouraging and discouraging. Encouraging because you could start to glimpse possibilities that EPUB will enable, and discouraging because the road ahead looks so steep.

Open Book Hack attracted a very stimulating mix of developers from around the world (who were in New York for Digital Book World) and local developers looking for interesting problems. Notably, there was a very impressive contingent of students from New York's Flatiron School. The projects are listed on a github wiki.

I was very eager to meet the folks from Berkeley's iSchool that have been working on epub.js, a javascript package that lets you read EPUBs from the web in your browser. One of them, Jake Hartnell, is also on the team building Hypothes.is, a tool designed to let everyone annotate the web (and one of the sponsors of the event). (The others were AJ and Fred.) Sooner or later you'll see these tools added to Unglue.it. But there's still a lot of work left to do.

For example, the project at Open Book Hack with the highest ratio of usefulness to impressive-sounding-ness was the project to add scrolling to epub.js (check it out, it works!). That's right, they added an option (mostly) to make chapter 2 come below chapter 1. You will soon be able to use ebooks on the web using an open-source 9th century reading interface.

There was one prize awarded. The winning project, "See and Read" allowed two people with two screens to interact via an ebook. Very nicely executed and well-deserved, but if I told you that I had invented something that allows two people to interact with a book, you would say "oh?" because you don't really need 100 billion transistors and two glowing screens to interact with a book.

Another project "Breadcrumbs and Beanstalks", extended Harvard's StackLife interface to enable ebook browsing similar to that found in a physical collection of books, with 2 dimensional browsing (the 2 axes being publication date and the general-to-specific axis of subject headings. It looked a bit clunky, but thar's gold in there somewhere.

Hugh MacGuire, Max Fenton, Jean Kaplansky, and Fendi M were doing something brave with linking in  PressBooks, but having already done too much linking in my day, I decided not to understand it. A team from Sophia University in Japan was doing something clever with textbooks and Readium. There was work on converting PDF to EPUB and a project to use Phonegap to make apps from EPUBs.

The project I joined up with focused on making a book club application around a shared EPUB reader. (It works, too!) We based it on epub.js and Hypothes.is. I wasn't very useful to the effort, because the three Flatiron-trained Ruby-on-Rails developers in our group were too awesome for my plodding-but-powerful python to compete with. So I helped with exploration and documentation of the Hypothes.is API, and finding bugs and deployment gotchas in epub.js. I now know how to configure CORS for buckets in S3. Yeah that was my weekend. I also fixed a bug by staring, in an intimidating way, over someone's shoulder. Ah, good times.

What became apparent to me in working with these tools was that the freshly trained developers got everything to work by un-EPUB-ing everything. The web platform just works, with the one exception being that centering text blocks in CSS just doesn't, unless you look away from the screen. The EPUB platform always throws something in your way, for reasons that even StackOverflow doesn't explain. Ruby Zips won't unzip. Cross Sites won't request. Java Scripts won't bind.

EPUB's competition isn't Amazon and KF8 fixed layout, it's the web and HTML5 and its huge gravitational pull. For 90% of ebooks, the benefits of EPUB over HTML are scant (because EPUB is based on HTML!) and the development barriers are significant. It's been years, and still EPUB authoring tools aren't mature or mass market. Deployment tools are barebones.

Don't get me wrong. I'm still betting big on EPUB, but dammit, Publishing Industry, for an $80 billion pillar of modern society, you're investing a nanoscopic amount on your basic infrastructure (i.e. EPUB), despite the herculean efforts of the people I met last weekend.

Notes:
  1. I was really impressed with the Flatiron School students I worked with. If the rest of the students are anything like Edina, Tiff, and Dan (Ivan helped a bit, too), they are going to have a huge impact on the New York area economy. Maybe I should learn RoR.
  2. Bill McCoy has done an amazing job bringing people together under the IDPF and Readium umbrellas. Imagine what he could do with financial support commensurate to his task. Perhaps he should take up bootlegging.
  3. Jake wrote up his impressions, too.
  4. As have Virginie Clayssen and Camille Pène, in French.
  5. Would have posted sooner, but MILESTONE IN UNLUE.IT.

Enhanced by Zemanta

0 comments:

Contribute a Comment