On interaction with data bases

Abstract

The main thesis of this paper is that interaction with data bases is fundamentally different from file manipulation: In the former case the data base should be "the master," while the file is a passive collection of data. The main corollaries of this observaction are: A) The conventional update and retrieval operations, traditionally used for files, are not satisfactory as the primitives of users data-base interaction. b) The privacy and integrity of the data base cannot be secured by "access control" alone; there must be a way for the data base to control the user's program itself.