RECOL(ID:186/rec006)REtrieval COmmand Languagefor REtrieval COmmand Language Climenson, RCA, Bethesda, Maryland, 1963 Early interrogative querying system - ran on RCA 501 5-stage interogative process:
Related languages
References: Introduction The primary reason for gathering data in some organized fashion is that a filing system provides a convenient reference to which one may go for information. An obvious requirement of a filing system is that the data must be accessible: some means must be available for pulling information out of the file. In some cases the retrieval task can be made more efficient with the use of data processing equipment. But machine storage of data files presents problems of (1) implementing an effective scheme for stating requests for information, and (2) defining an efficient file organization. The latter problem can be partially ignored by assuming that it is feasible to process a complete file serially--all file records are considered to be potentially relevant to a particular request. If serial processing is feasible and the concept of a file record can be used, the design of a file interrogation scheme is reduced to one of specifying and implementing an effective retrieval language. This paper describes a REtrieval COmmand Language (RECOL) which has sufficient flexibility for a number of applications requiring access to data stored in serial files. A program has been written for the RCA 501 to execute RECOL interrogation statements. It has been used successfully in several cases where it was desirable to implement a small or medium sized file interrogation problem on the computer with a minimum programming effort. It should be noted that specialized programs tailored to particular processing requirements are more efficient for complex tasks requiring periodic file manipulation or updating. No so-called "general" interrogation scheme would perform effectively in these cases. However, it is often desirable to mechanize files for short-term research or to handle a relatively inactive file as a special case within a large system; for such problems RECOL would provide early machine implementation at reasonable cost. Several other file interrogation schemes have been designed to include a flexible set of operations which can be controlled by the interrogator. Wang, for example, has developed a scheme which "turns out to be far more generalized than we originally intended". In the design of QUERY, the properties of natural language interrogation statements are used to characterize the retrieval function. VIP, though slanted toward manipulation of data (record rearrangement and combining), is another example of a generalized approach to file operations. in [ACM] CACM 6(03) (Mar 1963) view details One of the first identifiable data-base management systems to appear in the literature was an elegant generalized tape system developed by Climenson for the ROA 501 in 1962. This system, called Retfieval Command-Oriented Language, provided five basic commands, with Boolean statements permitted within some of them. The user had to specify the data description with the query so that a program could be bound to its data. in [ACM] ACM Computing Surveys (CSUR) 8(1) March 1976 view details |