Looking at computer-aided translation

As UNESCO's goal of education for all by 2015 looms ever closer, presenting an immediate and escalating requirement for content in a host of new languages, publishers and translators must find solutions that offer consistently high quality and high-volume scalability.

And with government increasingly moving towards e-content in education, publishers who remain stuck in the document formats and processes of yesteryear are finding themselves more and more at odds with the demands of their market.

In short, the new world of publishing is technology-enabled. Forward-thinking publishers and translators that embrace advances in quality, scalable content distribution on accessible platforms will be the flag bearers for a proud future in which everyone can participate.

Higher ground

Computer-aided translation (CAT) platforms are among our best hopes of hastening a future in which inclusive education and economic activity are substantial realities.

CAT occupies the high ground in-between human translation - which potentially offers high quality but cannot handle high volumes cost-effectively - and pure machine translation (e.g. Google Translate), which scales but offers inconsistent, below-average quality (Colina's assessment tool 2009, quoted in http://www.tandfonline.com/doi/abs/10.2989/16073614.2012.750824).

CAT combines human efforts with computing techniques to deliver quality and consistency that surpasses human and machine efforts, whilst also being faster than human translation.

Lack of awareness

Even so, awareness of what CAT can help publishers achieve is spotty. While the CAT portfolio of leading provider SDL, for example, is used by 80% of the global translation supply chain (including Rubric), CAT is not well understood in SA's publishing and corporate literature circles.

But with UNESCO's ambitious global goal of education for all by 2015, fast and efficient translation is becoming a focus area for educational publishers and not-for-profit translation services like African Network for Localisation. And as companies set their sights on transnational and emerging market growth, they too should research the CAT tools that can help broaden their markets.

How CAT streamlines translation

CAT tools use automation at selected stages of the translation continuum to streamline processes and iteratively enhance quality.

Traditional translation follows a translate-edit-proof approach. Subject matter experts write a text in Word (often in English), a translator translates it into Local Language 1 - also in Word - and sends it to the publisher. Here, a typesetter formats the translation in InDesign ('flowing'). This may take three - five days depending on the complexity of the layout. The work is printed to .pdf and sent to an editor, who makes corrections, sends it back to the typesetter, and a back-and-forth process ensues until both are satisfied. Implementing corrections may take from a few hours to several days depending on the number of corrections. This process has to be repeated for each additional local language.

Computer-aided translation cuts out the repetitive typesetting process for each language. The English document is translated in Word into all the local languages, using human translators. The translations are committed to a database, where phrases are stored with their translations for future reuse. In parallel, the typesetter flows the English document into InDesign, which, as noted above, takes three - five days. When the English document is typeset, it is speedily re-translated into all 10 local languages, using the database of phrases built up on the first pass. The translated documents are already pre-formatted in InDesign and the typesetter fixes the remaining minor formatting issues. This process takes a few hours, as opposed to 10 times that using the traditional method. Multiply that by 10 languages and the time saving is huge. For every update to the English text the CAT tool will help the translator stay in synch with the original. By the time the English original is approved, several iterations of the process have resulted in higher-quality typeset translations than would have been possible with the traditional approach.

By comparison, using the traditional method, each time the original English is changed, the editor, translator and typesetter have to iron out any disconnects between the English and the translations.

What it takes

So how can publishers and translators modify their businesses to get maximum benefit from CAT tools? In short, it requires a change in their processes. Users must accept that multiple parallel typesetting exercises are unnecessary.

But to typeset once for all documents and all subsequent edits, they will have to adopt a structured document format that sets out publishing text formatting and allows translated text to be updated or consumed in modular fashion, such as Extensible Markup Language (XML).

Get ready

The new world of publishing is technology-enabled. Publishers will have to avail themselves of the technologies that will enable greater efficiencies, but they must ready their own houses first.

The adoption of e-content in schools will help speed up the process by forcing the adoption of document standards - and may even have prevented the Limpopo textbook scandal. And in the end it will pay off for publishers too, as it will help improve sales and cut immense costs associated with printing and distribution.