A CDATA section is a region of text in an XML document delimited by <![[ at the beginning and ]]> at the end. All of the characters in a CDATA section are taken literally as part of the document. In other words, literal “<” characters, “&” characters, and other markup characters are not interpreted as markup. The delimiters are not rendered.

CDATA sections are a convenience for authoring, they are not part of the XML data model. After the document has been parsed, it is not possible to determine where CDATA sections where used (as opposed to other forms of escaping, such as character entities).

The intrinsic size of an image is it’s actual size, measured in pixels or other appropriate units, before any scaling or transformations have been applied.

Images in DocBook are defined by two bounding boxes: a bounding box for the overall image, the viewport area, and the bounding box for the image itself, the content area, which may be aligned within the viewport area.