Extensible, multi-presentation content is created in human / machine readable form, where each data object may contain numerous data elements that comprise alternate representations of a single piece of information, such as audio, text transcription of the audio, short-form summary of the text, and metadata concerning the story. Advantageously, each data element is presentation-independent, meaning that it does not require a particular type of presentation, such as HTML, WML, etc. Each content instance thus provides raw data, leaving the end-user free to access the data as desired. Specifications such as data type definitions (DTDS) may be used to prescribe that each story include the constituent data elements, e.g., text, audio, short-form text, metadata, etc. The DTDs define a dialect of extensible Markup Language (XML). Further, different DTDs prescribe the format of each data element. A developer may also define further DTDs defining additional data elements, sub-elements, sub-sub-elements, and the like.