The INTERNIST system, which is a computerbased diagnostic problem solver having Internal Medicine as its domain of discourse, employs a novel attention focusing heuristic in order to deal sequentially with the component parts of a complex clinical problem. The strategy of sequential problem formation and solution has proved to be effective in sorting out the complexities and rendering a correct diagnosis in the great majority of clinical cases tested. Experience with the system suggests, however, that many aspects of the search process could be significantly enhanced if it were possible to attend to the various component problems and their interrelationships simultaneously. Recent work on a successor system has been directed towards the development of strategies for the synthesis and analysis of composite hypotheses, which may be expected to yield more rapid convergence to the correct conclusion in many cases, and in at least some cases to prevent missed diagnoses. 1

Lisp is well known for its metacircular definitions. They differ by their intent (what they want to prove), their means (what linguistic features are allowed for the definition) and by their scope (what linguistic features are described). This paper provides a new metacircular definition for a complete Lisp system including traditionally neglected features such as cons, read, print and error. The programming style adopted for this interpreter is inspired both by denotational semantics and its continuation passing style (to explain continuation handling) and by the object oriented paradigm as highlighted by typedriven generic functions. The resulting interpreter lessens the number of primitives it uses to only eight: car, cdr, rplaca, rplacd, eq, read-char, write-char and end, while still providing Scheme-like essential capabilities (less arithmetic). The overall size is near 500 lines of fully encapsulated code that, if efficiency is not the main requirement, can be easily turned into...

. The Smalltalk programming language allows contexts (stack frames) to be accessed and manipulated in very general ways. This sometimes requires that contexts be retained even after they have terminated executing, and that they be reclaimed other than by LIFO stack discipline. The authoritative definition of Smalltalk [Goldberg and Robson 83] uses reference counting garbage collection to manage contexts, an approach found to be inadequate in practice [Krasner, et al. 83]. Deutsch and Schiffman have described a technique that uses an actual stack as much as possible [Deutsch and Schiffman 84]. Here we offer a less complex technique that we expect will have lower total overhead and reclaim many frames sooner and more easily. We are implementing our technique as part of a state of the art Smalltalk interpreter. The approach may apply to other languages that allow indefinite lifetimes for execution contexts, be they interpreted or compiled. 1 The Problem The Smalltalk language presents m...

In block structumed languages, upon entry to a block or procedure, storage is allocated for the identifiers declared in the block or procedure. There are two choices as to'when to deallocate this

This paper describee some new methods used in PLANNER-BESM-6 system for the implementation of PLANNER. Backtracking is implemented with a single stack and no copying of activation frames is needed. Lists are represented as arrays of pointers; the garbage collector for this representation requires no additional memory. Restrictions on variable values are accomplished by using 'semi-defined f structures. These methods have enabled to achieve the high efficiency of the system work. 1

An efficient backtracking method for LISP, used in the MLISP2 language, is described. The method is optimal in the following senses: to the desired goal. At each branching point in the tree, a decision must be made as to which alternative to try next. Backtracking is designed to simplify the programming of this type of problem. <1) Only necessary state information is saved. The backtracking system routines are sufficiently efficient to require less than ten percent of the execution time of typical jobs. (2) Most common operations — fetching/storing the value of a variable or the property of an atom, function entry/exit — take no longer with backtracking than without it. This is

The paper explains the principles of multi-result supercompilation. We introduce a formalism for representing supercompilation algorithms as rewriting rules for graphs of congurations. Some low-level technical details related to the implementation of multi-result supercompilation in MRSC are discussed. In particular, we consider the advantages of using spaghetti stacks for representing graphs of configurations.