關聯式資料群耦合程度衡量模式

Abstract

近來資料庫應用愈來愈受到重視，從階層式、網路式到關係式資料庫均受到企業界及學術界青睞，尤其是關聯式資料庫在近幾年更是凌駕於其他資料庫管理系統，但關聯式資料庫之規劃至今較為接受之方法在正規化及實體關係模式卻無-客觀之評量模式。 本論文研製之關聯式資料群之程度指標值，乃在於提供一客觀數值以供資料庫規劃、設計人員一重要之訊息，作為資料庫規劃設計之評量依據。本研究歸納出三種資料群間耦合之情形，並以推導之模式程度分別利用數學歸納法及代數法加以證明，繼而以該耦合程度衡量模式實際應用於企業之資料群，對於資料庫規劃之變動影響耦合程度指標值，並觀察其變化情形，探討耦合程度指標值之較佳範圍，此一耦合程度指標值之評估過程與結果，將能增進資料庫使用者與管理者對於各資料庫（群）間相關性質的了解，並提供往後衡量關聯式資料庫及分散式處理之研究基礎。

In recent years database applications have drawn much attention from both the academic circles and the business fields. The relational data model in particular has outstripped other models. 'The most widely used techniques for planning an ideal relational data base include normalization and entiey-relationship models, etc. Thus far, however, there is still a lack of design metrics for relational data bases. This research designs an index number for the degree-of-coupling measurement of ralational data groups. Three types of data groups are identified according to their coupling structures. The formul computing the degree for coupling is proven mathermatically. The physical meaning of the degree of coupling is illustrated by examples from simple to complex. The range of good coupling is observed and tradeoffs are discussed, This index number should help database designers and manegers characterize "coupling" strength in comparable ways and decide most desirable data group.