A Model and a Language for Representing and Manipulating Annotated Text Collections

Traditionally, collections of texts are digitally represented as a set of documents containing the text along with some kind of markup to define extra information, like metadata, annotations, etc. We propose a different approach that models the textual information in a dual way: as a formatted sequence of characters, as well as a composition of a particular kind of objects, called textual objects. With them, it is possible to represent different structures over the same text, together with complex annotations. Manuzio is a statically typechecked language to define a schema of such textual objects, and to write complex queries and applications on them with a set of powerful operators. In this paper we introduce the foundation of our textual model, the main features of the language, as well as a sketch of a system to manage persistent collections of texts and execute Manuzio programs.