臺灣博碩士論文加值系統

傳統文數字資料庫中查詢方式非常制式化,通常利用SQL作宣告式,成
批性查詢.在圖文資料庫環境下,圖形及文件資料常只是呆板的存放於關聯
表格中的某欄位中,查詢困難.自從標注技術(Markup)引進至文件中後,有
效建立了文件架構,使得查詢方式變得豐富.本研究試圖引進(圖面)架構於
任一圖形資料內,並推動圖形DTD(Drawing Type Definition,型態定義)方
式,限制,或是輔助繪圖工作,使得圖形內部架構,鏈結架構,屬性,及圖元等
資料結構,可形成有效之查詢路徑,允許使用者透過不同路徑取得所需圖文
資料.此外,由於圖形本身之特性適合人類協助處理,因此"以名查詢"和"瀏
覽查詢"兩種方式亦被包容於圖文查詢之可能方式之中.統合上述多種查詢
路徑,允許交錯程序性方式進行圖文一體之查詢模式被建立起來.基於前述
之圖文查詢理論,兩套驗證用的雛型系統,Static Query MAN ver2.5與
Dynamic Query MAN ver1.0,被發展出來.本研究搜集一群建築圖面及相關
文件,作為"實驗"內容,進行傳統查詢及多路徑交錯查詢之效能評估,據以
了解此種查詢模式之優越性.

Conventional alpha numerical database query methods are
limited togenerate declarative, high-level, set-oriented query
expressions, whichare not convenient for drawing data
retrievals. In our study, markup technique is introduced to
drawing data. A marked-up drawing data consists of four
fundamental structures, i.e.,hierarchy block relationship,
linking structure, attribute data, and"free-formatted" drawing
data. Based on these structures, users may finddifferent access
paths, such as get-by-navigation, get-by-retrievals, get-by-
hierarchy-traversals, or even get-by-manual-browsings. These
query methods are then integrated into a so-called Query Model,
which allows users to get drawing data, documents, and table
information together. In Static Query MAN, a primary query
path is embedded into this APsystem. In Dynamic Query MAN,
however, all query paths are integrated intothe Query Model.
Two prototyping systems, named Static Query MAN ver2.5, and
DynamicQuery MAN ver1.0, are developed. Finally, some
evaluation of various query paths based on different approaches,
such as SQL query method or the interactive query process
developed in this research, are discussed.