Brian Cantwell Smith

The implementation of procedurally reflective languages

In a procedurally reflective programming language, all programs are executed not through the agency of a primitive and inaccessible interpreter, but rather by the explicit running of a program that represents that interpreter. In the corresponding virtual machine, therefore, there are an infinite number of levels at which programs are processed, all simultaneously active. It is therefore a substantial question to show whether, and why, a reflective language is computationally tractable. We answer this question by showing how to produce an efficient implementation of a procedurally reflective language, based on the notion of a level-shifting processor. A series of general techniques, which should be applicable to reflective variants of any standard applicative or imperative programming languages, are illustrated in a complete implementation for a particular reflective LISP dialect called 3-LISP.

The owl and the electric encyclopedia

Abstract A review of “On the thresholds of knowledge”, by D.B. Lenat and E.A. Feigenbaum.

Limits of correctness in computers

On October 5, 1960, the American Ballistic Missile Early-Warning System station at Thule, Greenland, indicated a large contingent of Soviet missiles headed towards the United States [1]. Fortunately, common sense prevailed at the informal threat-assessment conference that was immediately convened: international tensions weren’t particularly high at the time, the system had only recently been installed, Kruschev was in New York, and all in all a massive Soviet attack seemed very unlikely. As a result no devastating counter-attack was launched. What was the problem? The moon had risen, and was reflecting radar signals back to earth. Needless to say, this lunar reflection hadn’t been predicted by the system’s designers.

Varieties of self-reference

The significance of any system of explicit representation depends not only on the immediate properties of its representational structures, but also on two aspects of the attendant circumstances: implicit relations among, and processes defined over, those individual representations, and larger circumstances in the world in which the whole representational system is embedded. This relativity of representation to circumstance facilitates local inference, and enables representation to connect with action, but it also limits expressive power, blocks generalisation, and inhibits communication. Thus there seems to be an inherent tension between the effectiveness of located action and the detachment of general-purpose reasoning. It is argued that various mechanisms of causally-connected self-reference enable a system to transcend the apparent tension, and partially escape the confines of circumstantial relativity. As well as examining self-reference in general, the paper shows how a variety of particular self-referential mechanisms --- autonymy, introspection, and reflection --- provide the means to overcome specific kinds of implicit relativity. These mechanisms are based on distinct notions of self: self as unity, self as complex system, self as independent agent. Their power derives from their ability to render explicit what would otherwise be implicit, and implicit what would otherwise be explicit, all the while maintaining causal connection between the two. Without this causal connection, a system would either be inexorably parochial, or else remain entirely disconnected from its subject matter. When appropriately connected, however, a self-referential system can move plastically back and forth between local effectiveness and detached generality.

Actor induction and meta-evaluation

The PLANNER project is continuing research in natural and effective means for embedding knowledge in procedures. In the course of this work we have succeeded in unifying the formalism around one fundamental concept: the ACTOR. Intuitively, an ACTOR is an active agent which plays a role on cue according to a script. We use the ACTOR metaphor to emphasize the inseparability of control and data flow in our model. Data structures, functions, semaphores, monitors, ports, descriptions, Quillian nets, logical formulae, numbers, identifiers, demons, processes, contexts, and data bases can all be shown to be special cases of actors. All of the above are objects with certain useful modes of behavior. Our formalism shows how all of these modes of behavior can be defined in terms of one kind of behavior: sending messages to actors. An actor is always invoked uniformly in exactly the same way regardless of whether it behaves as a recursive function, data structure, or process.

Quantitative description of the workload associated with airway management procedures

STUDY OBJECTIVES To measure the workload associated with specific airway management tasks. SETTING AND INTERVENTION Written survey instrument. PATIENTS 166 Stanford University and 75 University of California, San Diego, anesthesia providers. MEASUREMENTS AND MAIN RESULTS Subjects were asked to use a seven-point Likert-type scale to rate the level of perceived workload associated with different airway management tasks with respect to the physical effort, mental effort, and psychological stress they require to perform in the typical clinical setting. The 126 subjects completing questionnaires (overall 52% response rate) consisted of 43% faculty, 26% residents, 23% community practitioners, and 8% certified registered nurse-anesthetists (CRNAs). Faculty physicians generally scored lower workload measures than residents, whereas community practitioners had the highest workload scores. Overall, workload ratings were lowest for laryngeal mask airway (LMA) insertion and highest for awake fiberoptic intubation. Airway procedures performed on sleeping patients received lower workload ratings than comparable procedures performed on awake patients. Direct visualization procedures received lower workload ratings than fiberoptically guided procedures. CONCLUSIONS These kinds of data may permit more objective consideration of the nonmonetary costs of technical anesthesia procedures. The potential clinical benefits of the use of more complex airway management techniques may be partially offset by the impact of increased workload on other clinical demands.

Two lessons of logic

ceedings of the 7th International Joint Conference on Artificial Intelligence, Vancouver, B.C., pp. 331 -337. 1983. Natural language pmessing: crucible for computational theories of cognition. Proceedings of the 8th International Joint Conference on Artificial Intelligence, Karlsruhe. West Germany, pp. 1180-1186. SLAGLE. J R. 1971. Artificial intelligence: the heuristic programming approach. McGraw-Hill, New York, NY.

LINGUISTIC AND COMPUTATIONAL SEMANTICS

We argue that because the very concept of computation rests on notions of interpretation, the semantics of natural languages and the semantics of computational formalisms are in the deepest sense the same subject. The attempt to use computational formalisms in aid of an explanation of natural language semantics, therefore, is an enterprise that must be undertaken with particular care. We describe a framework for semantical analysis that we have used in the computational realm, and suggest that it may serve to underwrite computationally-oriented linguistic semantics as well. The major feature of this framework is the explicit recognition of both the declarative and the procedural import of meaningful expressions; we argue that whereas these two viewpoints have traditionally been taken as alternative, any comprehensive semantical theory must account for how both aspects of an expression contribute to its overall significance.