Many of the ideas behind object-oriented programming have roots going back to SIMULA. The first substantial interactive, display-based implementation was the SMALLTALK language. The object-oriented style has often been advocated for simulation programs, systems programming, graphics, and AI programming. The history of ideas has some additional threads including work on message passing as in ACTORS, and multiple inheritance as in FLAVORS. It is also related to a line of work in AI on the theory of frames and their implementation in knowledge representation languages such as KRL, KEE, FRL, and UNITS.

One day a man, who had lost Using an ordinary text-editing algorithm and a variety of much of his long-term episodic memory, consulted the professor changeable key words, the man could call up stories on his to ask him if there was any way he could help him personal computer, read them aloud, and thus attempt to regain the lost memories. Being righthanded text-editing method is trivial, but this is not an article and left-hemisphere specialized for language, he about method; it is about ethics.) The hope was that was still able to speak, to read and write: and to understand not only would the man now have some memory to think what was said to him. Besides the usual difficulty about and talk about but, more importantly, this repeated in recalling proper names, his main problem involved large daily practice at his own pace, with no one looking over gaps in his memory for events that he participated in before his shoulder, might help open up new access paths to his the stroke, although he could remember events that own memory of these events, filling them in and modifying occurred after the stroke. He could not, however, remember the award out the plan.

ECAI, Pisa 1984) to facilitate the decompiling of wealth and power.

During the balmy summer of 1980, about 800 AI researchers pose of the new format, the Committee's recommendation, met on the Stanford campus to hold the first and some expanded ways for members to participate in the AAAI conference. The conference program had no more conference this year. For many of Conference Goals those attendees, it was a special, unique opportunity to have deep colleagial interactions in a very comfortable setting. The most radical change that was considered, but not adopted, was the division of the science and engineering interests into two separate conferences at different times of Even the first national conference, however, was more the year. Many Council members expressed concern that than a gathering of researchers.

The major limitations in building large software have always been (a) its brittleness when confronted by problems that were not foreseen by its builders, and (by the amount of manpower required. The recent history of expert systems, for example highlights how constricting the brittleness and knowledge acquisition bottlenecks are. Moreover, standard software methodology (e.g., working from a detailed "spec") has proven of little use in AI, a field which by definition tackles ill- structured problems. How can these bottlenecks be widened? Attractive, elegant answers have included machine learning, automatic programming, and natural language understanding. But decades of work on such systems have convinced us that each of these approaches has difficulty "scaling up" for want a substantial base of real world knowledge.

When I wrote that we look for interest,ing articles, I made an exception, namely that we will not accept articles that reviewed or promoted commercial products. I felt that promotional articles were likely to be biased, sales-oriented and/or lacking in technical content. Reviews (particularly unfavorable ones) were bound to be criticized by the vendors as biased, uninformed, cursory, incomplete, etc. I still believe that these would be the results, although I have no evidence to prove it. At our annual AAAI Publications Committee meeting in August, the Committee suggested that I back off from this policy a little bit.

A variety of proposals for preferred methodological approaches has been advanced in the recent artificial intelligence (AI) literature. The article presents a review of such perspectives discussed in the existing literature and then considers a descriptive and relatively specific typology of these differing research perspectives. It is argued that researchers should make their methodological orientations explicit when communicating research results, to increase both the quality of research reports and their comprehensibility for other participants in the field. For a reader of the AI literature, an understanding of the various methodological perspectives will be of immediate benefit, giving a framework for understanding and evaluating research reports.

Further, new application domains such as computer -aided design (CAD), computer- aided manufacturing (CAM), and image understanding based on formal logic require novel concepts in knowledge representation and inference beyond the capabilities of current production rule systems. Fundamental research in artificial intelligence is concentrated at Corporate Research and Development (CR&D), with advanced development and applications pursued in parallel efforts by operating departments. The fundamental research and advanced applications activities are strongly coupled, providing research teams with opportunities for field evaluations of new concepts and systems. This article summarizes current research projects at CR&D and gives an overview of applications within the company.

This panel, which met in Austin, Texas, discussed the "deep unease among AI researchers who have been around more than the last four years or so ... that perhaps expectations about AI are too high, and that this will eventually result in disaster."

The choice of implication as a representation for empirical associations and for deduction as a model of inference requires a mechanism extraneous to deduction to manage uncertainty associated with inference. Consequently, the interpretation of representations of uncertainty is unclear. The calculation of representativeness depends on the nature of the associations between evidence and conclusions. We discuss an expert system that uses endorsements to control the search for the most representative conclusion, given evidence.