AQL(ID:1832/aql001)APL with picture primitivesPicture query language, extension of APL. Related languages
References: For expressing queries, the system we are investigating uses the subset of the natural language user's mother tongue) which relates to the particular universe of discourse (i.e. world's model) represented in the database. The meaning of this natural language formulation of the query is represented as a statement in a high-level non procedural language, which is, nevertheless, directly usable by the user who accepts a slight degree of formalization. The formal language provides a high degree of data independence in the sense that the underlying storage structures must not necessarily he known by the user, neither will his transactions the affected by changes in the storage structures or access strategies. Furthermore the formal language is embodied in an homogeneous environment in which the user may define programs and execute them on the results of queries. The memory model adapted is Codd's relational one: its symmetry and the semantic completeness of its algebra guarantee a friendly interface to the user who may look at the database as a set of named tables upon which he may execute the same (conceptually) operations he Is used to, i.e. table look-up, comparisons, etc. The Implementation guarantees data independence up to the level of changes in relation components or definitions. 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 in IBM Journal of Research and Development, 22(5) 1978 view details in Blaser, Albrecht (Ed.) Proceedings og the Conference on Data Base Techniques for Pictorial Applications, Florence, Italy, June 20-22, 1979, LNCS 81 Springer 1980, ISBN 3-540-09763-5 view details This paper presents a short but reasonably clear description of the author's query language, AQL. The examples are generally easy to follow, although at least one of them appears to be a little wrong. In discussing derived domains, GRANDFATHER OF PERSON is replaced by (FATHER OF PERSON) WHEN NAME ISONEOF (FATHER OF PERSON) which, if I understand the syntax and semantics correctly, would fail to find a maternal grandfather. A more serious objection has to do with the author's stated objective of providing "a relational query language intended for use by the non-computer specialist." AQL is less a query language than it is an extension of APL. For example, the authors suggest that a user might want to define a function "THEIR OWN" and proceed to write an APL function in APL syntax, implying that the user of AQL must also learn and program in APL. Whatever the merits of APL, I find it difficult to believe that even an extension of it, much less the full APL language, will be easy for a non-DP user, except possibly for the mathematically trained. I would certainly want to see some experimental evidence based on the experience of users with real problems to justify that position. The authors state that some applications are running on their system, but provide no other details. Until such results are available, it is probably best to suspend judgment as to the utility of this system for any non-APL user. F. J. Damerau, Yorktown Heights, N.Y. in ACM Computing Reviews 20(06) June 1979 view details |