A Translatours Catechisme (II): A New Drill

by Translation Guy on February 18, 2010

Written for the encouragement and instruction of all that have taken up the profession of translation, especially the common translator. Second in a series.

The ancient and honorable profession of translation is changing as never before. With the rise of the machine translators, and the technology that enables swarms of hobby translators (aka “community translators”), the profession is beset by opportunity and challenge.

As the field of translation expands exponentially, growth is fastest at the bottom of the market. Why should people pay for translation if they can get it instantly for free, or by persuading others to contribute their translation labors for nothing? Sometimes it seems as if the profession of translation is beset on all sides, and a new translation model is required if professional translators are to create the value that will secure their slice of the translation pie.

The old professional translation model: translation, edit, proof (TEP) — sometimes observed, sometimes given lip service only — hardly distinguishes a professional workflow from the work of the amateurs who spill their text over the pages of Facebook and the video frames of Japanese manga.

So professionals, both client and vendor, would do well to apply a New Model of Translation to their work. Language service providers (LSPs) such as my own, 1-800-Translate, will profit themselves and their customers using New Model Translation, as we transform from LSPs to language management providers.

So to answer this call, we promulgate a new translation doctrine. Here’s the drill to replace TEP with the handy acronym AGPTEPEPVHU.

A is for Align: Gather up all the translations that have been done  in the past and push them through  an alignment tool which identifies each phrase and matches it to its translation, creating a translation memory database and identifying terms for the glossary.

G is for Glossary Construction: Now we have a handy list of critical words (which may have unique meaning in a particular context) that we can then make our translators use, so they won’t be fishing around in a dictionary getting creative.

P is for Pre-translation: We run the materials to be translated against the TMX database, match up all the previously translated phrases with the new project, and send the partially translated file to the translator, which allows them to work much more accurately and efficiently.

T is for Translation: ‘Nuff said…everybody knows what that means.

E is for Edit: The 11 categories of translation error get corrected at this point.

P is for Proof: Now the corrections can get corrected too.

E is for Export: Using translation memory, it is easy to push this content into any format or software automatically, this being done with greater efficiency and accuracy than any programming or production John Henry could ever manage. So what clients once did heroically on their own is now in the domain of the New Model Translator.

P is for Proof, again: Never trust a machine, or the operator behind it.

V is for Validation: How do you know if your New Model Translator is on the right track? Sometimes we know we don’t know, and sometimes we don’t know what we don’t know. (I love that line.) Others (distributors, in-country people, focus groups, in-house language management) may know best, or may only be able to comment on a message after it is translated. Lots of ways to do this.

H is for Harmonization: That’s the part when “trust but verify your validators” comes into play.

U is for Update: Make sure everything you learned is incorporated into the translation system so the same misteaks (!) are not repeated.

Sounds complicated, huh? But that’s the new drill for the New Model Translator. Fortunately, technology makes it a lot easier than it sounds and produces a professional product easily distinguished from the works of machines and hobbyist with whom we must share the field.

We know from our training and testing work for P&G and others that this is a critical requirement for translation in your industry, which is why we are recommending this workflow to you. It could have a big impact on price, too.


  1. ForensicCult says:

    My translation servicess make a point to say that all content passes through a complete TEP cycle. In your opinion Ken, is this enough?

    • Ken says:

      Is TEP enough? It all depends on what you are doing with it. If the translation is for outreach, as in marketing or ad copy, simple TEP leads to broken hearts and in-country dissapointment more often than not. But for something inbound, who cares. But for enterprise language managment, its an essential starting point. Its easier to drop steps in a workflow when you don’t need them than to add them when you do.

  2. Katie says:

    If historical growth and acquisition trends hold, who you buy translation and localization from in 2013 could be quite different from who you buy it from today (although the sales rep may still be the same and have the same phone number)

  3. How does this new ‘model’ or your proposed model, have such an impact on price…can you elaborate on that briefly?

  4. Chuck says:

    As a society, we are at the very early stages of the emergence of a new model for translation of online content – “peer production” models of translation. Thoughts//?

  5. Olivia says:

    I think this blog has covered this in some form or another, just wondering if there were any thoughts on this becoming a realistic means of quality translation?

    • Ken says:

      Olivia, getting beyond TEP is not just realistic, but is the new professional translation reality. We apply it whenever we can. But for one-off or smaller projects, steps are dropped. For our customers I would say the validation/harmonization piece is the hardest part to apply. Budgets and resources are usually not available. They have to rely on us to get it right the first time. Do you think I am being unrealistic, Olivia?

LiveZilla Live Chat Software