Localising graphics

Where possible, graphics that include text should be managed in SVG format to simplify localisation.

In many cases, documents written in a DITA environment will be translated to another language. Localisation software programs and XML languages work incredibly well together, so the localisation process is generally very efficient. However, localisation runs into a stumbling block where graphics include words.

The problem is essentially eliminated if the graphic is in Scaleable Vector Graphics format, because, being an XML format, translatable text can be extracted out of an image and the translated text reinserted.

For raster graphics, where the text in a graphic is made up of uneditable pixels, you should use a key for callouts. The callout therefore becomes a key reference (for example, A) to the description, which is typically recorded in a table or definition list under the graphic.