The industry's concern for quality results in improved translation protocols.
The translation industry has experienced an interesting evolution. In the beginning, clients were translating individual documents in a single language. Then, the demand for multiple languages increased, along with the need to handle multiple projects for the same client. Then came translation memory, content management, and single sourcing.
We are now at a level of sophistication where the need to look at the translation process and how it fits into a larger quality system is important. This is a daunting task, given the complexity of (any) human language, and the subjective elements that constitute it. This article will: look at the most important current models being used and developed around the world, review the status of the standard being developed by the automotive industry, and review steps that the translation consumer can take to ensure that his/her translation is of good quality.
When reviewing current standards, the Germans have already developed a standard relating specifically to the translation process. This standard, called DIN 2345, is similar to ISO standards in that it concentrates more on the process of translation than on evaluating the linguistic elements that are produced. The DIN standard addresses the following issues: the original text, choosing the right translator, setting up a contract between the client and translator, the target text (how well it conforms to requirements of that target language), and proofreading (what to look for). Proponents of the DIN standard say that in contrast to ISO, this standard allows flexibility to meet the different demands and output for different clients. It offers a framework that allows you to change the system without changing the quality of the final product.
From a development standpoint, there are currently two major efforts to develop quality standards. The first, proposed by the ComitŽ EuropŽen de Normalisation (CEN), is the result of the combining and editing of norms submitted by 17 different countries. In the U.S., the Association of Language Companies (ALC) has participated in the development process. The work, which is being supervised by the British Standards Institute, will be considered by ISO for approval as its international standard for the industry. This document will focus on the process of translation, rather than on evaluating the quality of the text produced by a translation team.
Within the U.S., ASTM is currently in the final editing stages of its Standard Guide to Quality Assurance in Translation and Localization. As with CEN's document, this document provides guidelines for the process that should be agreed upon by the client and translation provider to produce a final translated product of the desired quality.
Unlike these two efforts, the translation needs of the automotive industry in the U.S. have led to the development of a Recommended Practice to address this industry's concern for quality. Developed in conjunction with the Society of Automotive Engineers, the J2450 defines a quality metric as it applies to technical automotive documentation. The recommended practice attempts to define quality while avoiding the difficult stylistic issues that any type of written content poses. For this reason, the use of the J2450 standard is limited to technical material. It represents one of the first, if not the only, standard to attempt to measure the quality of the translated product in a codified way.
SAE, working in conjunction with representatives from General Motors, Ford and Chrysler, along with representatives from the translation industry, came together in 1997 to develop a metric for the evaluation of translation quality that could be used by the entire automotive industry and its translation suppliers.
Due to the huge volume of translation required by automotive manufacturers for international (and now domestic) markets, they have spent a great deal of time and money developing a system that works well for them. It relies on the use of language evaluators to look at a translated document and rate the number and type of errors in that document. The practice involves classifying errors into categories and sub-categories, and then computing a weighted score corresponding to the quality of the translation.
A challenge posed by this system is its implementation. The evaluation must be done by a native speaker, and that native speaker must be very familiar with the system in order to produce reliable results. In addition, if you are producing the translation in several languages, making sure that the evaluators are similarly trained has a great deal to do with making sure that the results are consistent across languages for a single project.
There are a few additional challenges posed by the J2450. One of these is that although there is a statistical method for evaluating errors found in a translation, the results are dependent on the number of words in the document. There must then be a way for determining whether these results are acceptable for a document of this size. There are no currently defined guidelines as to what results represent an acceptable translated document. This would have to be developed by each individual company for its own needs.
While these represent major trends in the translation industry, the question of quality still starts with the translation consumer, and his or her relationship with a translation provider. To help make sure that you get good, consistent translation quality, there a few basic steps.
Carefully select the appropriate translation provider for your projects. Make sure that you define your needs and audience before starting to work with a particular company, and make sure that they can meet those needs, both on a short-term as well as a long-term basis.
Working closely with that provider, develop tools such as glossaries and translation memories. A glossary of terms that provides both the term and a good definition will help ensure that your material is well understood. This glossary should include not only terms specific to your industry, but also specific to your company. Include common abbreviations, and any company jargon which might be unfamiliar to someone outside your company. The use of translation memory software could be an asset if you have a good deal of repetition in your documents, a fair amount of documentation, and a file format that interfaces well with the translation memory software being used.
Make sure that your source language is as clear and concise as possible, avoiding any possible misinterpretations of the original. Provide any necessary background or product material. Remember that a translation can only be as good as (but is sometimes better than) the original English. Instill in your writing staff the importance of consistency and a concise presentation. It makes the English more usable, and it becomes easier to translate.
In addition, you can develop your own quality control system that involves reviews of the translated documents. Using internal and in-country resources can help you get an evaluation of how appropriate the translated document is to your target audience. This feedback must be shared with your translation provider, as it is critical to getting a good translated product in the long-term. Remember that your translation provider needs to learn about your style, preferences and terminology. Until there is a consistent feedback "loop" the translator cannot change and improve his or her output.
Working well with your translation provider can provide huge quality benefits. Due to the challenges (especially linguistic) in developing an accepted translation standard, it will take some time to be able to look for this type of assistance in quality efforts. Until something is available, remember that a consistent process, while not guaranteeing an excellent final product, goes a long way in ensuring quality work. Translation is a collaborative process involving a number of people. Keeping these concepts in mind can help you achieve your personal quality goals for the translated product.