Diacritical Support: The Whole World, and Nook
I’m evaluating an EPUB project that is seemingly pretty straightforward, if long. It’s a dictionary, 1700 pages, 85,000+ entries.
In looking through the pages, I realized that all those diacriticals that help pronunciation, including popular items like the schwa, the bilabial implosive, and the bilabial trill, might pose a problem for some EPUB readers.
I was confident that iBooks and KF8 would do well, but what about the Nook? This book will be sold through Barnes and Noble (it’s a major market for this publisher), so it must “work” perfectly.
A colleague provided a link to the International Phonetic Alphabet in Unicode, and I created a simple EPUB3 document. Here’s a look at part of the markup:
Here’s a screengrab of the iBooks (iOS9, iPhone) version:
The iPad is just as responsible; it displays all elements. I opened the file in the Kobo app on my iPad, and all’s well there.
I threw the EPUB into Kindle Previewer for a quick test there, and all seems fine.
But the Nook. What about the Nook? Well, from the Nook app on my iPad:
See those boxes? The first 3 items are not displayed. The rest of the book follows suit: some displayed, many not. And it’s the same for the Nook HD, the Simple Touch, and the iPhone app.
Even the simple and ubiquitous schwa (ə further down the list) gets boxed out; no display.
I went to Twitter, and @ helped by saying that the Nook runs Adobe RMSDK (the Adobe toolkit for EPUBs), which does not support these characters. She also mentioned a similar problem with Arabic. (As a side note, I launched the book in Adobe Reader 4, and was happy to see all my symbols. Tzviya tells me that ADE4 is based on Readium, which you can read about here.)
A Font Solution?
Keith Snyder (a fine detective and epubsecrets.com contributor; @noteon) suggested embedding a font.
So I went looking, and came across this writeup on The Linguistic Mystic, which led me to a font, Charis, which I downloaded. An open-source, true type font. I loaded it into my EPUB, enclosed the entities in a span calling on the embedded font, and here’s the result:
Now I’ll submit the EPUB to the editorial team to make sure all needed diacriticals display properly, and I’ll send the license to the lawyers to make sure we’re allowed to use it. There’s nothing like some crowd-sourced solution building!
There might well be other issues with this huge book, including working from a complex XML document, and a gigantic number of cross-references, requiring massive scripting and perhaps risking markup overload. More to come.