Detailed discussion of bar‐hillel's “theoretical aspects of the mechanization of literature searching”

Abstract

AbstractBar‐Hillel is criticized as tending to use the terms “impossible” or “unfeasible” for almost anything that cannot be done immediately. This is considered inappropriate for evaluating research. It is stated that there seems to be no reason why a lower limit in the costs of machine processing should exist that would be above comparable costs of human processing. This is presented to counter Bar‐Hillel's doubts as to whether machines can do complex information processing tasks (in the area of documentation) cheaper than present methods. A “proof” that machines are capable of fully‐automatic, high‐quality indexing, extracting, abstracting, etc., is presented to refute Bar‐Hillel's doubts on this point (and as a counter‐proof to Bar‐Hillel's demonstration of the nonfeasibility of fully‐automatic, high‐quality machine translation of languages). Other topics discussed include: types of information retrieval systems and distinctions between various complex information processing tasks.