Multi-language e-Books

Multi-language Background

Our experience with multi-language textbook production and delivery has been an eye-opener to the possibilities of multi-language production and format delivery across the full book production spectrum.

Digital content delivery has instantly liberated multi-language publishing strategies. For example not only can we produce textbooks in multiple languages but in ethno-linguistic minority languages where there has never been any print material available (and never will be).

But it doesn't stop there. The textbooks we have produced are multi-language interactive aural textbooks. That is, they read the text and animate the content where appropriate. In-house we call it the blackboard mode. But this article is about multi-language pubishing in general rather than our recent textbook experiences.

Multi-language books (who have never multi-languaged before) open new business channel opportunities for publishers of all types.

First... Translation

Translation is not trivial and the required skills for different types of content vary widely. For example translating a novel is arguably the most challenging translation work. Technical and textbooks are a little more straight-forward.

However our job was not to translate but rather make the process of translation of formalnonbreakingspace book content easier and more flexible with the required review checks and balances. The other challenge was including multiple languages in the same book (to date we have maxed out at four parallel languages).

The IGP:Digital Publisher International Language Extension (ILE) opens a new way to translate formal publisher content. There is no root language defined. Through the language extensions a translator can translate from any available language to their target language. Translation is paragraph by paragraph and can be reviewed and commented at any time by an authorized reviewer.

Remember it's not just about textbooks. The IGP:Digital Publisher ILE works for books, documents and content of all types.

The translation decision map

We had to make a decision on translation granularity carried out by people in locations around the world. After a number of trials and attempts at short cuts we settled on parallel paragraphs as being the most powerful and useful approach. (That includes headings, list items, table cells and other content containing elements.) That meant containing blocks (sections and divs, etc.) were language neutral. This approach has worked out well.

Each paragraph has the lang attribute with the appropriate language value. In addition each paragraph has the same ID with a language suffix to separate them. Eg: The IDs for four paragraphs in four languages look like this:

<p id="Ea001-en" lang="en">...</p>
<p id="Ea001-hi" lang="hi">...</p>
<p id="Ea001-mr" lang="mr">...</p>
<p id="Ea001-ur" lang="ur">...</p>

The advantage of this approach is that we can process and add interactivity by content item ID and by language. If a target language needs multiple paragraphs in translation they can simply be added with an ID letter extension. Eg: Ea001a-hi

The tranlation tools allow a translator to select source and target languages and get to work and not have to worry about viewing any other languages.

The Multi-language Reading Systems

We have packaged the books using ePub3 using the advanced features of the AZARDI:Content Fulfilment and the AZARDI ePub3 reading application. The Javascript language selector is currently packaged with each book and is now an official component in the AZARDI Interactive Engine (AIE).

The books are designed to work with full interactivity on Mobile phones down to five-inch screens; specifically Android 4.4.1 and above phablets and tablets. They also work on all desktop operating systems. We haven't looked at Apple phones or tablets as they are hardly a presence in rural environments in developing countries although AZARDI does work on iOS with both a WebApp and native app available.

This exercise highlighted the significant limitations of the Epub3 specification and the arbitrary ePub4.1 ePubCheck "upgrade" (five years after the specification was released). We have been able to work around that but are waiting with bated breath for the Web Platform Working Group to get their work on the ground so digital book content is freed from these arbitrary backward looking specifications.

Another critical issue is book metadata. Currently there is no mechanism to allow the book title to be presented in the user's reading system language of choice so we have to use polyglot covers and custom metadata that only AZARDI Reading Systems will be able to currently understand.

Multi-language production

Once tranlation has been done the editorial and production team gets to work. They can be working on one or multiple languages, even while another language is being added by a translator.

We have re-engineered IGP:Digital Publisher  and IGP:FoundationXHTML to handle multiple languages with relative ease. We are fortunate that internally we have staff who have multiple language skills in Indic and West European languages. That makes the development of new tools and process approaches a lot easier as we have a team of internal critics. 

In addition to parallel paragraph translation, production and presentation options, multiple languages can be used in SVG illustrations and audio files. This subject will be dealt with in a series of separate articles.

Publishers take note...

If you haven't already worked it out, digital content radically changes a publisher's multi-language options and potentially open new markets around the world. You can address new language markets without going to the cost of print.

Translated books  can be released in many languages, including parallel language editions or  exclusive language digital editions. If print editions are required IGP:Digital Publisher allows any layout and combination of PDF editions to be created either for print on demand or market-specific print editions.

The costs and challenges of printing and physical distribution for translations have been significantly reduced in this digital content production and delivery era. But there are a number of new challenges for multi-language production and distribution.

With the publisher content production power of IGP:Digital Publisher it is entirely realistic to have any book translated and reviewed and release it as a set of digital editions in multiple languages.

We don't expect trade publishers to jump on the international band-wagon immediately, but education publishers working in markets with multiple languages should look at it very seriously right now.

Posted by Richard Pipe