Practical Relational Approach
Because textual data tends to occur in repeating groups, it does not lend itself well to normalized database structures. Frequently, the use of such structures results in performance and database maintenance problems which complicate and compromise data reliability and availability.
In the library world, for example, the MARC record comprises hundreds of repeatable fields, each subfield divided into large numbers of repeatable components. Normalization of these records presents tremendous challenges, not only at record assembly time, but also for the database administrator who must maintain the tables. Similarly, in the world of Electronic Data Interchange (EDI), transaction data are arranged in repeating groups, with the potential for large numbers of such groups in individual messages and, more importantly, in the databases from which those messages are constructed. Rather than forcing normalization upon database designers, MINISIS supports it, as well as supporting repeating groups within tables. Fields can be added to, removed from, or relocated within groups and tables as required. The result is improved performance and maintainability.