SURGE: A Syntactic Realization Grammar for Text Generation
FUF and SURGE
This is the home page of the FUF and SURGE packages. FUF is an interpreter
written in CommonLisp for a functional unification based language
specifically designed to develop text generation applications. SURGE is a
comprehensive generation grammar of English written in FUF.
FUF has been developed by
SURGE has been developed by Michael Elhadad and
The packages are available under the terms of the
SURGE requires FUF. FUF is written in Common Lisp. The interpreter should
work on any Common-Lisp. It has been tested on Lucid Common Lisp versions
2 to 4.0 on Sun 386, Sun3, and Sun4s, DEC stations and HP workstations.
Allegro Common Lisp has been tested on Sun3s and Sun4 (sparcstations) and
on NeXt machines. Allegro Common Lisp has been tested on the Macintosh II
and PC Windows version 2 and 3.
(Contact me for specific
instructions about the PC version.) Ibuki Common Lisp has been tested on
Suns. FUF has also been tested under the POPLOG version of Common Lisp.
It should really be compiled to work honestly. Tail recursion elimination
must be enabled to produce best results.
What is in the package:
The FUF package contains:
Full Common Lisp source for the interpreter.
Complete documentation (200 page postscript user manual).
A tutorial on how to write generation grammars using FUF.
The SURGE package contains:
Full FUF source of the syntactic realization grammar.
Regression testing system with approx 500 examples of English
The following papers and dissertations contain in-depth descriptions of
FUF/SURGE related material:
Controlling Content Realization with Functional Unification Grammars,
Michael Elhadad and Jacques Robin, in Aspects of Automated Natural
Language Generation, R.Dale, E.Hovy, D.Rosner and O.Stock eds,
Springer Verlag, 1992, pp89-104
An Overview of SURGE: A reusable comprehensive syntactic realization
component, 10 pages,
Floating constraints in Lexical Choice,
Michael Elhadad, Kathleen McKeown and Jacques Robin,
to appear in Computational Linguistics, 1996, 40 pages,
Using argumentation to control lexical choice: a unification-based
PhD thesis, Columbia University, Dept of Computer Science, 1993,
Revision-based generation of natural language summaries providing
historical background: corpus-based analysis, design, implementation
PhD thesis, Columbia University, Dept of Computer Science, 1994,