State-of-the-art language industry standards mirror best practices in industrial standards. Reflecting a transition that has been widely completed in the manufacturing and services sector, language-related QA standards are undergoing an evolution from the notion of quality control to that of quality assurance. Within this context, translation and localization quality are judged on the basis of text "quality," however it may be defined for the localization of mostly commercial products. The distinction between process and product, which has been a part of theoretical debate in translation studies for decades, aligns serendipitously with the process/product approach of the core ISO 9000 quality system. Finally, this article will consider current industry efforts to establish process standards and product metrics, such as the DIN, ÖNORM and ASTM guidelines governing contract negotiations for translation and localization services, and the SAE metric for translation quality.

1. Quality control vs. quality assurance

ISO 9000:2000 states that quality assurance (QA) provides confidence that quality requirements will be fulfilled, whereas quality control (QC) focuses on checking quality product and services. QA encompasses procedures set in place to assure and document quality throughout the production process, whereas QC in the localization industry has historically involved end-item inspection, in the form of editing, proofreading, and product testing. The LISA QA Model sums up its QA orientation as "an ongoing process [that] should not happen just at the end of a project" .

Although this distinction between QA and QC is perhaps more honored in the breach than in actual practice, it reflects historical development from QC via QA to the current concept of total quality management (TQM) and provides the structure for the standards hierarchy.

2. Typology for pragmatic translation in localization environments

Schnitzlein (2003; 2 ff.) observes that there is all too often a disconnect between practice, research (pragmatic translation studies), and training, not to mention a further disconnect between approaches used in industry and government. There is actually good reason for industry's relative disregard for translation theory insofar as theory has focused solely on literary values or tried to transfer sometimes obtuse philosophical concerns to discussions on pragmatic translation. Truth to tell, much of this effort is unproductive with respect to localization. Office workers sitting down to use a new word processing software in Sri Lanka or Singapore are not interested in perceiving the socio-cultural essence of the source text shining through a "foreignized" translation - they only want to get on with their tasks, using their own languages and counting on the program to support them with the appropriate character sets, currency values, formatting conventions, etc. Regardless which of the more or less "standard" programs they use, the programming interface itself comprises its own cultural environment, which should at this point be roughly international, stripped of any references to source language or culture.

Nevertheless, there are some theoretical views that are useful for evaluating quality in localization environments.