Abstract: Contemporary database query processing systems base their actions principally on "syntactic" considerations, and seek only the most efficacious way of answering a query as originally formulated. An alternative approach seeks to use knowledge of the semantics of the database's application to transform the original query into an alternative form, possibly quite different in its expression, but which is both equivalent to the original (in terms of the set of records from the database that it qualifies) and more efficient to process, given the existing file structures and access methods. The architecture of a system supporting such knowledge-based "semantic" transformations has been developed. It addresses such issues as the kinds of knowledge that should be included in the knowledge base and how it should be expressed, the kinds of transformations that can exploit this knowledge to improve query processing, and the way in which the system as a whole can be organized in the presence of large and intricate knowledge bases and a multiplicity of possible transformation types. This latter structure is based on a multi-processing model, in which each possible transformation is treated as a process, whose priority is assigned by a scheduler embodying a variety of heuristics. The principal contribution of the work is the establishment of a conceptual framework for this type of query optimization and the design of an architecture that can grow with the development of additional transformation techniques.

Abstract: In this chapter, we present techniques that allow a query system to play an active (or intelligent) role in communicating knowledge in large databases to the user via meaningful and efficient feedback during query execution. Our approach is to dynamically create temporary files and access paths for information relevant to present query and inform the user of the existence of such information (if security is not compromised). Our research and experience shows that such information can be made accessible at a very low cost when the system is obtaining the data that has been requested and can be presented (if user shows interest) to the user without much effort. Four types of semantics (database semantics, database organization semantics, usage semantics, and real-world semantics) have been identified, their sources, appropriate data structures for their representation, and applicability have been presented with examples from medical and pictorial databases. Our hypothesis for this research is that most users cannot be expected to know all necessary information available in a large database and the query system must play an intelligent role and provide hints to the users.

Abstract: An online system of associative information retrieval is described. The system uses a network representation of the information in a bibliographic database and a processing paradigm modeling a continuing flow of user interest through the network to implement associative retrieval and query modification. The system is capable of relevance feedback, thesaurus and statistical query expansion, Boolean and best‐match searching, and retention of associations based on previous search experience. The query modification capability stems from the ability to incorporate thesaurus or dictionary linkages between terms and preferred or alternative forms and to generate associations between sets of documents and the vocabulary they contain. The former capability allows the automatic incorporation in a query of pre‐defined equivalent or alternate terms. The use of terms associated with retrieved documents allows automatic discovery and inclusion of related vocabulary, index terms, and subject categories and may improve the effectiveness of free text searching in databases incorporating controlled vocabulary indexing and subject classification schemes. A pilot version of the system has been implemented on the DECsystem‐10 and tested on small scale files;illustrative sample searches are presented.

Abstract: In [4], Lipski proposed a mathematical model of data bases with incomplete information and discussed some problems related to it. According to him, propositions which express queries to an information storage and retrieval system can be regarded as a special kind of formulas of the first-order predicate logic. So, in [3] he gave two way ( i.e, external and internal ) of interpreting formulas of the predicate logic, by using models of data bases with incomplete information. In regard to internal interpretations, some similarities to Kripke models for modal logics are known. In fact, certain relationships to the modal logic $4 were mentioned in [3]. In this paper, we will show some results on internal interpretations for a query language. In Section 2, a translation of formulas in a query language, called extended formulas, into formulas in the secondorder language is introduced and then it will be shown that an extended formula is true in every internal interpretation if and only if the corresponding second-order formula is true in every second-order interpretation. By using this fact, we will prove in Section 3 recursive solvability of some decision problems related to a query language for data bases with incomplete information. As a corollary, we can answer affirmatively a conjecture proposed in [3], which says that internal equivalence of extended formulas containing only monadic predicate symbols is decidable.

Abstract: : The existing human factors literature on query language is both sparse and scattered. This paper seeks to collect and review that literature. The first section of the paper introduces the subject of query languages. In the second and third sections, the topics of natural and formal query languages are respectively discussed. These two types of query languages are reviewed with the objective of determining their potential for expanding the population of computer users. The fourth section considers some general issues pertinent to both types of query languages. These issues include the ability of people to deal with logical quantifiers, the user's concept of data organization, mixed initiative dialogues, and the use of abbreviations. Methods for experimentally evaluating specific query language features and research on person-to-person communication are also discussed here. To focus the findings reported in the preceding sections, the fifth section summarizes the implications of the research performed to date. Next, the sixth section presents possible new research which would be of value to the designers of Army tactical information systems. The paper concludes with two appendixes. Appendix A discusses human factors review papers concerned with the design of interactive systems. Appendix B presents a compendium of design recommendations directed towards the system designer.