The Fnc-2 Attribute Grammar System

Didier PARIGOT

Oscar Project

FNC-2 is a modern AG-processing system that aims at production-quality by providing the following qualities:

external Efficiency
the generated evaluators, based on the visit-sequence paradigm, are completely deterministic; furthermore, this paradigm allows to apply very effective space optimization techniques. The evaluators are hence basically as efficient in time and space as hand-written programs using a tree as internal data structure.
external Expressive power
this efficiency is not achieved at the expense of expressive power since FNC-2 accepts AGs in the very broad class of strongly non-circular AGs. Explore new formalisms allowing more comfortable use of AGs (generic AGs )[ LeBellec93, LeBellec93a Roussel95] or more expressive power (Dynamic AGs [Parigot96a] , which subsume circular AGs and allow to work on DAGs and maybe more general graphs).
external Easiness of use
OLGA, the input language of FNC-2, enforces a high degree of programming safety, reliability and productivity. The input language containt the following sub-languages:
external Versatility
the generated evaluators can be interfaced with many other tools and produced in many variations

external Some applications
external Companion processors
In addition, FNC-2 comes with a number of companion processors that help to build complete applications

external How to get the FNC-2 system

external Publications
The principal articles on FNC-2 attribute grammar are:


Web page maintained by Didier Parigot
Wed Mar 11 12:58:19 MET 1998