A body section contains the main elements where localizable or translatable texts are kept. It is called the 'translation unit' or <trans-

unit› element in the XLIFF format file. It contains an identity, 'id' attribute, to map where a segment is located in the source language text. Figure 2 illustrates this. The trans-unit element has a source element, which is the segment that is to be translated, and a target element, which is the accepted translated segment.

Figure 2 Example of a header in XLIFF

Figure 3 Example of a body in XLIFF

Sometimes, a second translation element, 'alt-trans', is given as an alternative translation for the source segment, as shown in Figure 4. An alt-trans element also contains a source and a target segment. This element can be matched from a translation memory, machine translation system or be suggested by a translator.

In Figure 4, the alt-trans element is matched from a translation database of a translation memory system. Using specific matching algorithms, an alternative translation is found but the matching, which is not perfect, only measures at about 60 per cent (match-quality). This indicates that the alt-trans element is a fuzzy match at best. Alt-trans is also a way of keeping track of the changes made during the stages of the localization process, a useful mechanism for keeping a record of what

Figure 4 Example of an alternate translation element in XLIFF

the changes were, who did them, which tools were used to make them, and so on. Once the translation work has been completed, the XLIFF format file is reconverted into the original file format (HTML) and the non-translatable portion (layout) of the source-

language text is reincorporated. An XLIFF format contains only one source language and one target language. It is different from the TMX format that is able to support many languages in the same document.

责任编辑：admin