The growing importance of language industry standards for reusing – or 'leveraging' - data previously processed by other tools cannot be underestimated in the modern world of the translation industry. These standards function as a bridge, connecting tools that are built using different programming codes and operating across a variety of platforms or systems, such as Microsoft Windows or Unix. The existence of numerous tools has made it necessary for specific standards to be created to help the operational flow when several tools are used concurrently in a translation process. Tables 1 and 2 focus on this aspect but from different perspectives.

Table 1 shows how important standards are for each translation type, referring to three different types of standard: TMX, TBX and

XLIFF. Within machine translation, the need for particular standards may vary according to the sub-type. So, whereas direct translation and rule-based machine translation systems do not use previously translated material or corpora to generate new translations, corpus-based machine translation systems do have a use for data interchange standards, as do also computer-aided translation systems, including some linguistic tools such as electronic dictionaries and termbases. At present, however, the TMX standard for the interchange of translation data is applicable mostly to translation memory systems as they store previously translated material in a database of translation examples, rather than to full machine translation systems.

Table 1 Data interchange standards in translation

VI=very important; I=important; P=possible

We can recall that machine translation systems are increasingly being used to generate the first drafts of translations in the localization process. In such cases, the output data from the machine translation system – the raw translation – needs to be imported into the localization tool for further processing. For this to happen, the XLIFF standard is needed. The terminological data interchange standard TBX, on the other hand, is extremely important to all translation types, especially for the translation of technical texts. For computer-aided translation and human translation, both translation and terminological data interchange are very important, especially when linguistic tools or resources such as electronic dictionaries, glossaries and thesauri are used to assist the translator during the translation process, since most are compiled and structured in different ways. Furthermore, they may be accessed via different operating systems. The localization standard XLIFF is important for computer-aided translation and human translators; however, its importance is clearly restricted to supporting translators involved in the localization industry.