next up previous
Next: Story Data Up: ESPER: architecture Previous: Identifying the Speakers


In our work we have designed an XML markup language so that we can mark information within our original text preserving its layout. Current we have only used the elements CSML, QUOTE and CHARACTER. QUOTE elements support TYPE and SPEAKER attributes. CHARACTER attributes are ID (whose values are the same as SPEAKER in QUOTES, and CLASS (which describes the syntactic type of the character name). We will expand this support as necessary.

Alan W Black 2003-10-20