for QUEry Analysis SYstem
introductory query language
The influence of human factors on achieving the optimal design of a management information system is presently receiving considerable emphasis. Specifically, the relationship between the syntax of the query language and the ease with which a user acquires a familiarity with the skills used in the manipulation of a data base is being investigated.
In a recent paper, Thomas and Gould [i] attempted to evaluate the effectiveness of a natural language format used by IBM's SEQUAL  as contrasted to a more formalized approach used by Zloff's QUERY BY EXAMPLE . The results of their experiments indicate that the use of a formalized system for the query language format seems to benefit the type of casual user as described by Codd , principally in terms of the ease of learning and by providing cues that result in greater retnetion of acquired skills. As a side effect of this formalization, the usefulness of the system for more advanced users diminishes since the query structures are limited.
The requirements of this present research was to design a management information system that could achieve a simultaneous enhancement of skill effectiveness for both casual and sophisticated users. The Query Analysis System (QUEASY) was developed to satisfy these requirements.
The specifications of the system required support for three levels of user'skills and retrieval requirements. The system's lowest, but largest level contains the casual users. These casual users are characterized as having limited time for skill acquisition, minimal need for retraining sessions, controlled access to the data, and possibly only a short term effectiveness. The mid-level of users are responsible for the integrity of the data provided by casual users, and the subsequent authorization of the data entry. This userlevel will have and require more time for initial training in the use of the query language, but will function for longer periods of time. The functions of the highest user-level are akin to the functions of the data base administrator. These individuals will be fully trained in the manipulation language in order to conduct statistical processing of the data base, and they will have complete authority over the authorization of data access and entry procedures. Extract: Methodology
The following decisions were the primary factors which influenced the design and implementation of the QUEASY system:
1. The underlying data base structure should be the relational model.
2. The design should be done in a topdown structured manner.
3. The system should be available in a timesharing environment.
4. QUERY BY EXAMPLE should be used as the model for the query language.
Extract: System Design
In the evaluation of the first design decision, the choice was from among the three existing data base structures: hierarchical, network and relational. Given the size of the data base, the large number of casual users and the need for a relatively sophisticated user interface, the relational structure was the best for this application. The primary advantage of the relational structure is that the user is spared from having to know anything about the underlying physical structure of the data. The primary disadvantage is that a query on a relational data base may be extremely inefficient.
Since the relational structure is well documented in other sources, this paper will not attempt any formal definitions. In simplistic terms a relational data base can be viewed as a collection of tables, called relations, which contain entries, called tuples. In the other data base structures a relation is roughly equivalent to a record type and a tuple is roughly equivalent to a logical record.
For the second design decision, QUEASY was designed using structured top-down procedures. The system HIPO (Hierarchy plus Input-Process-Output) Chart is given in Figure i. This chart gives the major programs that were developed for the QUEASY system. Due to extremely severe memory limitations on the computer resources available for the project, it was decided to break the system down into these separate, relatively small, programs.
In response to the third design decision, the system was designed for a timesharing environment where the majority of users were computer-naive. To assist this type of user'several features, such as an extensive HELP command, prompting cues for user input, giving a reasonable response at any time to a question mark and relative flexibility in data input, were included in the design. Also, special data types were developed to aid the user in interpretation of the data, such as PASS and FAIL being displayed instead of a less meaningful number representing these two values.
The fourth design decision was a direct result of the literature review. From the available query language models, QUERY BY EXAMPLE was the best for use by casual users. QUERY BY EXAMPLE is a graphics oriented query language developed by IBM for a relational data base system. Since the project had no interactive graphics capabilities, it was decided to implement the important features of QUERY BY EXAMPLE without interactive graphics. Extract: System Implementation
The actual coding of the QUEASY system was done in standard FORTRAN IV for the Control Data Corporation Cyber 173 operating under the NOS-I operating system. This computer is used by the Chancellor's Office of the California State Universities and Colleges, which is located in Los Angeles, California. The important programs of the system all require less than 28K words of memory for this computer.
During the course of this project there were two main implementation phases: system testing and user training. During the spring of 1978 the system was tested using standard structured techniques for a trial data base. Operational tests are scheduled to begin during late spring 1978 for a new experimental self-paced Mathematics course.
in Proceedings of the 1978 annual conference 1978, Washington, D.C., United States view details