It may seem counterintuitive, but audio is a primary medium for providing accessibility to many persons with disabilities. Whether it is an audio recording of someone reading text, or a realtime computer generated "Text To Speech (TTS)," rendition, persons who are blind, or who live with severly impaired vision or a learning disabilities, often use audio as their primary reading modality. While we do not expect ODF applications should become audio recording and editing applications, there are nevertheless critical considerations that should be observed in order that documents produced by ODF applications might easily be used to create the smart audio renditions increasingly used by people who rely on audio renditions of textual content. 5.1 Where and How Audio Is Used for Accessibility Audio renditions of textual content are so common and powerful that they have been codified in an ANSI/NISO standard, Z39.86. This same specification has been adopted internationally by a consortium of libraries for the blind and print handicapped called the DAISY Consortium. In turn, this ANSI specification served as the basis for the U.S. legal mandate to provide accessible text books and curricular material in U.S. Schools, known as NIMAS. Materials produced in audio include everything from novels for leisure reading, to newspapers, magazines, and technical reference material (in addition to curricular material). There are also national programs for creating and distributing such content across Europe, Canada, Australia, and Japan, as well as many other countries. 5.2 How ODF Fits In ODF authoring applications are relevant to alternative media production, including audio, simply because alternative media such as audio renditions are just that--a different media edition of the same content produced for visual consumption by ODF applications: Item: Soft Page Breaksand Hard Page Numbering Ref to spec revision][Ref to spec revision] Even the simplest content cannot be discussed accessibility in a group environment if there is no simple mechanism to point group members to a particular location in the document. Paginagion is the most common resolution to this problem,, however pagination is inextricably tied to a particular rendring of content. Alternative media such as audio (and large print and braille) must have mechanisms to allow their users to know where they are relative to the hard page numbers of the primary source document. Item: Structural Markup Effective use of audio renditions requires that users have the ability to move quickly back and forth through the audio rendition based on the structure of the document. Traditional audio playback equipment provided fast forward and rewind mechanisms, but these are highly inefficient because time offsets are actually irrelevant to content. What is relevant is the structure of the document? Does it have Chapters? Sub Sections? Footnotes? Sidebars? Paragraphs? Effective support of alternative rendering, and especially audio rendering, requires that the source document be correctly tagged with structural markup. Indeed, the aforementioned ANSI and NIMAS specifications provide XML based markup to allow rendering agents to support quick movement forward and backward through content based on chapters, subsections, footnotes, paragraphs, and other structural elements. The most usable devices allow users to adjust "levels" of navigation, so that hierarchical structures such as X.Y.Z might be navigated at the X level, the Y level, or the Z level, at the user's option.