A Programming Language for Finite State Transducers

SFST-PL is a programming language for finite-state transducers which is based on extended regular expressions with variables. SFST-PL is used by the Stuttgart Finite-State-Transducer (SFST) tools which are available under the GNU public license. SFST-PL was designed as a general programming language for the development of tokenizers, pattern recognizers, computational morphologies and other FST applications. The first SFST application was the SMOR morphology [1], a large-scale German morphology which covers composition, derivation and inflection. An SFST program consists of a list of variable and alphabet assignments followed by a single regular expression which defines the resulting transducer.