Finite models and the theory of concatenation

We propose FC, a logic on words that combines the previous approaches of finite-model theory and the theory of concatenation, and that has immediate applications in information extraction and database theory in the form of document spanners. Like the theory of concatenation, FC is built around word equations; in contrast to it, its semantics are defined to only allow finite models, by limiting the universe to a word and all its subwords. As a consequence of this, FC has many of the desirable properties of FO[<], while being far more expressive. Most noteworthy among these desirable properties are sufficient criteria for efficient model checking and capturing various complexity classes by extending the logic with appropriate closure or iteration operators. These results allows us to obtain new insights into and techniques for the expressive power and efficient evaluation of document spanners. In fact, FC provides us with a general framework for reasoning about words that has potential applications far beyond document spanners.