Nordic Journal of Computing AML: ATTRIBUTE GRAMMARS IN ML Sofoklis G. Efremidis Department of Computer Science Cornell University Ithaca, New York, USA sofoklis@cs.cornell.edu Khalid A. Mughal Department of Informatics University of Bergen N-5020 Bergen, Norway khalid@ii.uib.no John H. Reppy AT&T Labs Research 600 Mountain Ave. Murray Hill, New Jersey, USA jhr@research.att.com Lars Søraas Department of Informatics University of Bergen N-5020 Bergen, Norway larss@ii.uib.no Abstract. Attribute grammars are a valuable tool for constructing compilers and building user interfaces. This paper reports on a system we are developing, called AML (for Attribution in ML), which is an attribute grammar toolkit for building such applications as language-based programming environments using SML. This system builds on the proven technology of efficient attribute evaluation, while using a higher-level foundation for the implementation of interactive systems. It supports a general and uniform platform for building applications that can manipulate attributed terms and allow access to attribute values. We describe the design of the AML system, its current implementation status, and our plans for the future. CR Classification: D.1.1, D.2.6, D.3.4 Key words: attribute grammars, attribute evaluation, functional programming, program generator, programming environments 1. Introduction Attribute grammars provide a formalism for assigning meaning to parse trees of a context-free language [26]. Because of their syntax-directed form and declarative style, they provide a useful notation for specifying compilers [23] and language- based editors [37]. This paper reports on a system we are developing, called AML (for Attribution in ML), which is an attribute grammar toolkit for building applications such as language-based editors using SML [28]. This work was supported, in part, DARPA-ONR Grant N00014–91–J–4123. Current address: Intracom S.A., 190 02 Peania Attica, Greece. E-mail: sefr@intranet.gr Support for this work was provided, in part, by the Norwegian Research Council. Current address: NERA Telecommunications ASA, Kokstadveien 12, N-5061 Kokstad, Norway. E-mail: las@nera.no Received July 1996. Revised December 1996. Accepted January 1997.