Well, it can entail a lot of overhead, or none at all; that depends on the needs and inclinations of the user. Users creating content can totally ignore it. Content creators who have multilingual content, that they want to encode as multilingual content, are “required” (which means encouraged) to include one metadata attribute (@lang-grouping) to flag that they are using this mechanism. In addition to the @lang-grouping attribute, users are expected to use as much, and only as much, of this infrastructure as is appropriate to their content and expected use. The minimum would be tagging all language groups with the @lang-group and @xml:lang attributes. Content receivers are not required to accept documents that use the multi-language mechanism, and can easily identify documents that use it tagging by looking for the @lang-grouping flag on the <processing-meta> element. Content receivers that do accept and process multilingual documents using the JATS multi-language mechanism can use the @lang-grouping flag to identify those documents that need this additional level of processing.