If semi-structured documents are implemented, it would be nice to have a way to use text from the document itself to fill the structured data. Like the YummySoup “any site web importer” works for recipes. You import text and mark blocks of text as title, ingredients, directions, etc, then import the recipe and those fields are filled appropriately. So, for semi-strucured documents, it would be really cool if I could just import a document, select a block of text, and mark it as author, date, journal or whatever, and DT would then fill those fields with my choices.