Downloads & Free Reading Options - Results

A simulation study of library-based information retrieval systems is described. Basic models for each of several important aspects are presented: (1) user behavior, emphasizing response to quality and delays in services; (2) the scheduling of services and the organization of the machine-readable files; and (3) the distribution of conventional library materials. Many of the variables in the model (e.g., user expectations of service time, delays involved in providing services, etc.) are random. An adaptive element is present in the form of admitting changes in user expectations in response to system changes. The need for frequent model change and the merging of component models into a comprehensive model for the entire system is established as the basis for use of a user-oriented simulation language. Statistical facility beyond that supplied in the language used (GPSS/360) is required to achieve the goals of this and related investigations. (Author)

This experiment assessed the impact of using two levels of retrieval terms on formulating and inputting query statements. One level used specific terms to retrieve one element of information; the other level used global terms to retrieve blocks of information that would otherwise require five to six specific terms. Participants were then given a set of 48 problems. For each problem, the participants had to write and type a query statement that would satisfy the information requirements. The opportunity to use global terms had not effect either on the time needed to write query statements or on the accuracy of typed query statements. Where the use of global terms was applicable, substantial savings in the time required to input query statements was shown. Except that the global-specific group reported that it made more use of the data name chart before using the data dictionary, the two groups indicated that they went about writing query statements in approximately the same way. Both groups gave high ratings to the value of using global terms. Use of global terms is not recommended unless the specific items of information subsumed under the global term are normally retrieved together frequently.

This experiment assessed the impact of using two levels of retrieval terms on formulating and inputting query statements. One level used specific terms to retrieve one element of information; the other level used global terms to retrieve blocks of information that would otherwise require five to six specific terms. Participants were then given a set of 48 problems. For each problem, the participants had to write and type a query statement that would satisfy the information requirements. The opportunity to use global terms had not effect either on the time needed to write query statements or on the accuracy of typed query statements. Where the use of global terms was applicable, substantial savings in the time required to input query statements was shown. Except that the global-specific group reported that it made more use of the data name chart before using the data dictionary, the two groups indicated that they went about writing query statements in approximately the same way. Both groups gave high ratings to the value of using global terms. Use of global terms is not recommended unless the specific items of information subsumed under the global term are normally retrieved together frequently.