DITA comes with several built-in features to help create global content. One good feature is that DITA topics are stored in Unicode, so codepage problems are a thing of the past. This allows you to use the same output devices for every language, without worrying about which is supported.
You will generally want to set the current language on the root element of a DITA file. This language is then used to process each item in the topic; any automatic headings will be generated using that language. The language is also used for formatting information. For example, Hebrew and Arabic will both use right-to-left formatting rather than left-to-right. Additionally, captions for Hungarian tables appear in a different order than most other supported languages, with the table or figure number appearing before the "Table" or "Figure" text.
The language setting is actually available on nearly every element in a DITA topic. This allows you to create multiple language documents, where each element is processed based on the currently specified language. For example, you could create a document with Danger notices in various languages. You will need to get the warning itself translated, but the generated "Danger" title will already show up correctly. Note that for this to work, your chosen output format must support multiple languages -- if you plan to generate a correctly sorted index, you are not going to have much luck with an index that spans many languages.
Another of DITA's useful features is the translate attribute. If you have an item that should not be translated, set the translate attribute to "no". Human translators and translation tools will automatically skip this value, saving you time and cost in the translation process. This is especially useful when creating specialized DTDs, as described below.
