Lolli -- An Interpreter for Linear-Logic Programming In conjunction with the Joint International Conference and Symposium on Logic Programming, and in particular the Workshop on Linear Logic and Logic Programming, we are pleased to announce the first public distribution of Lolli, an interpreter for logic programming based on linear logic principles. The language Lolli (named for the linear logic implication operator -o, called lollipop), is a complete implementation of the language described in the paper ``Logic Programming in a Fragment of Intuitionistic Linear Logic" (Joshua S. Hodas & Dale Miller, LICS '91, an extended version to appear in Information and Computation), though it differs a bit in syntax, and includes a wide variety of built-in extra-logical predicates. The logic underlying Lolli can be viewed as a refinement of the the Hereditary Harrop formulas of Lambda-Prolog. All the operators (though not the higher order unification) of Lambda-Prolog are supported, but with the addition of linear variations. Thus a Lolli program distinguishes between clauses that can be used as many, or as few, times as desired, and those that must be used exactly once. These features have been used to provide a perspicuous, logical implementation of a number of example problems, including object-oriented programming with mutable state and information hiding, database update, and term rewriting. Lolli has also been used as the setting for the development of a filler-gap dependency parser for natural language processing. This last is described in the paper ``Specifying Filler-Gap Dependency Parsers in a Linear Logic Programming Language", by Joshua S. Hodas, to be presented at JICSLP '92. This preliminary implementation was developed over the last year and is based on code written by Frank Pfenning and Conal Elliot for their paper ``A Semi-Functional Implementation of a Higher-Order Logic Programming Language" which appears in ``Topics in Advanced Language Implementation", MIT Press, Peter Lee editor. DVI files for this paper, as well as the two papers mentioned above, are included with the distribution. Obtaining the Lolli Distribution ================================ To retrieve a copy of the Lolli system, ftp (anonymously) to ftp.cis.upenn.edu (130.91.6.8) and retrieve the file pub/Lolli/Lolli-0701.tar.Z. (Be sure to set ftp to BINARY transfer first.) This distribution includes full ML source, along with a Makefile, as well as the papers mentioned above and a collection of example programs. If you retrieve the system, please send e-mail to hodas@saul.cis.upenn.edu so that you may be kept informed of updates. For those who do not have SML-NJ 0.75 at their site, we hope to provide pre-built binaries for a variety of architectures. These binaries can be found on ftp.cis.upenn.edu in the directory pub/Lolli/binaries. At present Sparc and NeXT binaries are available. If you compile Lolli on a new architecture, please contact us so that we can make your binary available.