基礎時空圖徵架構之發展與應用探討

Development and Application of Primitive Spatio-Temporal Feature Framework

由使用者觀點，地理資訊系統應用之重要關鍵為可正確掌握各類資料之時空狀態，尤其在以圖徵為基礎之架構下，正確之時空現象描述是資料供應單位無可推卻的責任。然而在開放資料供應之層次上，時空現象之描述卻常因缺乏一致之設計策略，造成不同單位所供應之圖徵設計架構各自不同及整合運作之困難。本文以單一圖徵為基礎，提出其基本架構必須至少包含識別碼、時間、空間描述與主題屬性等四類不同之元素，並進一步依時間及位置變化之型態而具體歸納四類基礎之圖徵架構，可因應不同領域由簡單到複雜的時空現象描述需求。每一類基礎圖徵架構均由前述之四項元素組成，且因應其時空變化之型態而設計符合其特性之內部架構，可避免重複及無謂之資料流通。當基礎於如GML 之開放資料格式而發展實作機制時，應用系統發展者將可因此以固定方式剖析參考基礎圖徵架構而設計之各類圖徵資料，達成減低應用模組開發成本及促進資料開放與互操作性之目標。參考基礎圖徵架構而供應之圖徵將可協助使用者端完整掌握其時空因素，並正確及有效地加以整合。 From users’ perspective, one of the key issues of GIS applications is to build a correctunderstanding about the spatio-temporal property of the acquired data. Particularly for thefeature-based applications, it is an unshirkable duty for the data providers to supply correctspatio-temporal descriptions about the phenomena in reality. Due to the lack of a consistentstrategy for feature design, the open frameworks of geospatial data from different domains areoften quite different and such heterogeneity may cause unpredictable difficulties during dataintegration. This paper argues that the fundamental components of a feature must at least includeidentifier, time, location and attributes, and further proposes four types of feature frameworksdependent on how the location and attributes of the described phenomena change. These four typesof frameworks enable designers of different professional domains to describe from simple tocomplicated spatio-temporal phenomena in reality. Every primitive feature framework iscomposed of these four components. Their inner structures are designed in a way that theduplicated and unnecessary data recordings can be effectively avoided. When implemented with open data formats like GML, application developers can thus consistently parse the feature dataaccording to the type of framework it refers, which help to attain the goal of reducing the cost ofmodule development and facilitate the open and interoperable data use. By referring to theprimitive feature framework, users can build a complete understanding towards thespatio-temporal status of the acquired data, and can thus correctly and effectively integrate the datain their applications.

From users’ perspective, one of the key issues of GIS applications is to build a correctunderstanding about the spatio-temporal property of the acquired data. Particularly for thefeature-based applications, it is an unshirkable duty for the data providers to supply correctspatio-temporal descriptions about the phenomena in reality. Due to the lack of a consistentstrategy for feature design, the open frameworks of geospatial data from different domains areoften quite different and such heterogeneity may cause unpredictable difficulties during dataintegration. This paper argues that the fundamental components of a feature must at least includeidentifier, time, location and attributes, and further proposes four types of feature frameworksdependent on how the location and attributes of the described phenomena change. These four typesof frameworks enable designers of different professional domains to describe from simple tocomplicated spatio-temporal phenomena in reality. Every primitive feature framework iscomposed of these four components. Their inner structures are designed in a way that theduplicated and unnecessary data recordings can be effectively avoided. When implemented with open data formats like GML, application developers can thus consistently parse the feature dataaccording to the type of framework it refers, which help to attain the goal of reducing the cost ofmodule development and facilitate the open and interoperable data use. By referring to theprimitive feature framework, users can build a complete understanding towards thespatio-temporal status of the acquired data, and can thus correctly and effectively integrate the datain their applications.