TAMALAN(ID:4118/tam002)? for the Tree? Van Dyck, CDC, 1977 Relational data retrieval language, interfacing to CODASYL database system (first to do so) Query language for Comm/Sci, uses Command, Rel. calculus Dev: CDC (Belg) HW: CDC Structures: References: in Information Systems 2(4) 1977 view details in Proceedings of the 3rd conference on very large data bases (Tokyo, Japan, Oct. 6-8, 1977), IEEE, New York, ACM, New York, 1977 view details This paper introduces a different approach to defining query languages called "domain oriented." Domains are defined independently of relations, e.g., a COMPANY domain might be ("AMC," "CHRYSLER," "FORD," "GM," "VW"). The relations become ways of building associations between the entities of the domains. Many query languages such as, for example, SQL, SQUARE,. TAMALAN have a tuple-oriented approach. This approach introduces tuple variables and name qualification within a tuple. The domain-oriented approach avoids this but involves heavier quantification. Based on the domain-oriented approach and a domain relational calculus, an "English-like" language called ILL (Intermediate Level Language) is defined. Care has been taken to guarantee that ILL is "wholly built on a structure of expressions nested inside other expressions...." This is, of course, generally useful to assure expressive power and flexibility, independent of any English-like attribute. The hypothesis of the paper is that the domain-oriented approach has a closer similarity to English than the tuple-oriented approach. This is supported by a number of examples which contrast ILL with the tuple-oriented language SEQUEL (now called SQL). Certain constructs are not included in ILL in order to retain its English-like quality. This seems unfortunate. A construct can frequently be useful without being English-like. It is, or should be, well known that some queries are difficult to express in English, particularly those in which negation and quantification interact. The ILL language becomes somewhat complicated here as well. Actually, both tuple- and domain-oriented approaches are useful in different circumstances. Perhaps what is needed is a query language that permits a natural integration of both approaches. In any event, the domain-oriented approach is a nice way of handling many queries. This paper presents the idea in a clear and well organized way. It is well worth studying. D. Lomet, Yorktown Heights, N. Y. in ACM Computing Reviews 20(09) September 1979 view details types. IBM's Structural English Query Language (SQL), developed for use on a relational DBMS, is also examined and compared with RIDL. Conceptual and usage differences are discussed and similarities assessed. in ACM Computing Reviews 20(09) September 1979 view details |