臺灣博碩士論文加值系統

資料庫法則式系統 (Database Rule System) 是藉著監視資料庫狀態
(state) 以達成自發性 (active) 資料庫的一種方法, 意即,當資料庫系
統在操作的過程中, 系統要轉變到一個法則預設的狀態,則此對應的法則
便啟動去更改這種轉變。法則式系統除了能達成自發性資料庫功能外, 它
還能用來實作或加強這些資料庫的功能, 以一種更一般化或彈性化的方式
。大多數的資料庫法則式系統是在查詢語言 (query language)的環境中
來做, 或者是在查詢語言上架構一層邏輯式的語法來達成。在本篇論文
中, 我們嘗試提出一個新的法則式的模型 (model) ,它是建立在資料庫的
程式語言 (programming language) 的環境中。我們的模型是將法則
(rule) 含入資料個體 (object) 中,因此,它整合屬性 (attribute)、法
則 (rule)、和路徑 (method)個體內路徑的行為 (behavior) 必須受到法
則的限制, 因為法則代表知識 (knowledge) 能力, 所以 我們把這種個體
資料結構叫做知識個體 (knowledge object)。我們的法則式結構可以以
事件- 條件- 動作 (Event-Condition-Action)來描述: 當事件發生, 引
發 (trigger) 法則, 條件被檢查;當條件成立, 則執行動作。 我們的模
型實作在ONTOS物件導向資料庫上,在此, 事件相關的對象是個體的路徑,
條件和動作則各為使用者所指明的任何限制 (constraint) 條件和操作
(operation)。

Database rule system is a mechanism which monitors the database
state} to make the database active. That is, when the database
is changed into an expected state, the rule system triggers an
appropriate rule to take further action change. Besides
supporting the active database applications, the rule system is
useful for implementing several functions of the database
itself, often in a more general and flexible way than provided
by current implementations. Most database rule systems are
written in a query language or in a logic language. In this
thesis, we try to design a rule system written in the
programming language. In our rule model, rules are incorporated
into objects. Therefore, an object contains attributes,
methods, and rules, and the behavior of a method must be
constrained by the associated rules. This kind of objects are
called knowledge objects in the sense that rules are kind of
knowledge. Our rule structure has the following Event-Condition-
Action (ECA) semantics: when an event occurs, it triggers a
rule; then the condition of a rule is checked; if the condition
is satisfied, the action of a rule are performed. Our model is
implemented on ONTOS object-oriented database system. In the
implementation, an event is associated with a method, whereas
condition and action could be any user-defined constraint and
operation, respectively.