In the sections below we consider three major text types which present specific problems: literary texts, office documents, and language corpora and other collections. Language corpora are a major genre of machine-readable texts. By and large they exhibit the same sorts of problems as traditional collections of other kinds, such as newspapers/periodicals and anthologies. In office documents special attention must be paid to front matter (addresser, addressee, version, etc.). Genre variation and authorial creativity cause problems in the markup of literary texts.