How Multi-language Documents are Presented and Stored

Both the presentation styles and the order of storing components are widely variable for current multi-language documents. When substantial portions of document content are in more than one language, the same structures are often repeated, once in each language. For example: alternate sections could repeat the same content in Spanish and Portuguese. As another, a paragraph could be repeated, first in Greek, then in Romanian, and then in Italian. Sometimes an article will be almost entirely in one language with selected structures in both (or only) another language, as article written in Korean with both metadata and table content provided in English. [[JATSMultilang]]

In such a multilingual document, when there are two or more same-content structures (sections, figures, boxed-texts, tables, etc.), differing only in language, few assumptions can be made about the locations of and the interrelationships between these same-content objects within the document.

For these reasons, a JATS multi-language mechanism could not simply enclose (wrap up) all the same-content objects, unless the entire article was presented, in order, twice. This wrapper-approach has been the mechanism of the long-standing JATS element <block-alternatives>, which was created to hold multiple copies of a block object such as a figure or table in multiple languages. Users tried to use <block-alternatives> for multilingual documents and found that it did not meet real-life needs. For true interspersed multilingualism, a wrapper-style mechanism is not sufficient. [[JATSMultilang]]