以領域樣版促進高效能TWSMP詮釋資料建置之研究－－以國土利用資料為例

Using Domain-specific Template to Improve the Efficiency for Generating TWSMP Metadata：An Example of Landuse Data

詮釋資料為國家空間資源流通與共享機制之必要元件。為有效提升詮釋資料管理及查詢之效能，普遍之作法為制訂詮釋資料標準，進而統一約制各單位詮釋資料之建置內容。標準雖然提供一致遵循之規定，但複雜之引用技術及專業之填寫規定往往需要大量訓練，形成全面推廣之障礙。詮釋資料樣版（Template）雖非嶄新之概念，但如何正確建立合適之樣版卻缺乏具體之策略及評估之成果。本文所提出之基本策略為經由縝密之分析而建立資料特性與詮釋資料項目之對應關係，再由專家依分析資料之現況而規劃專業樣版，由樣版之引入而彌補建置人員專業之不足。以所提出之三階段建置程序實際測試國土利用調查資料之詮釋資料建置，顯示納入領域之專業樣版後，不但描述內容之專業性及完整性可有效改善，需由建置人員處理之詮釋資料項目也大幅縮減至總建置項目之14%，充分彰顯使用樣版之成效。相對於目前基層建置人員因訓練及專業不足而僅能填寫有限詮釋資料項目之窘境，本文之研究成果具有重大之指標意義。若能具體落實，未來之詮釋資料建置任務將轉換為由上而下之整體配套規劃，全面提升運作之效能及描述內容之完整性。 Metadata is a necessary component to the distribution and sharing of national geospatialresources. To effectively improve the efficiency for managing and querying metadata, a commonapproach is to use metadata standard to constrain the metadata content created by differentorganizations. Although standards can provide a set of consistent rules to follow, the complicatedencoding technology and professional requirement on entry content often require a tremendousvolume of training and thus impede the generation of metadata. Metadata template is not a new idea, but how to create appropriate template still lack of a comprehensive strategy and good practice. Weargued that the development of template requires to first analyze the corresponding relationshipbetween the property of geospatial data and metadata elements, then asks domain experts to createtemplate according to the status of geospatial data, such template can then be used to compensate theprofessional deficiency of metadata creator. By applying the proposed 3-stage metadata generationprocedure to the national landuse data, we demonstrate that not only the quality of metadata contentis largely improved, the number of required processing metadata elements is also reduced to only14% of the total metadata elements. The domain-specific template clearly demonstrates itsadvantages on improving the efficiency for generating metadata. Compared with the currentsituation, where the generation of metadata is restricted to a limited number element due to the lackof training and professional knowledge, this paper points out a possible direction to resolve thisobstacle. The future metadata generation will thus be based on a top-down approach that can bothimprove the efficiency for generation and quality of contents.

Metadata is a necessary component to the distribution and sharing of national geospatialresources. To effectively improve the efficiency for managing and querying metadata, a commonapproach is to use metadata standard to constrain the metadata content created by differentorganizations. Although standards can provide a set of consistent rules to follow, the complicatedencoding technology and professional requirement on entry content often require a tremendousvolume of training and thus impede the generation of metadata. Metadata template is not a new idea, but how to create appropriate template still lack of a comprehensive strategy and good practice. Weargued that the development of template requires to first analyze the corresponding relationshipbetween the property of geospatial data and metadata elements, then asks domain experts to createtemplate according to the status of geospatial data, such template can then be used to compensate theprofessional deficiency of metadata creator. By applying the proposed 3-stage metadata generationprocedure to the national landuse data, we demonstrate that not only the quality of metadata contentis largely improved, the number of required processing metadata elements is also reduced to only14% of the total metadata elements. The domain-specific template clearly demonstrates itsadvantages on improving the efficiency for generating metadata. Compared with the currentsituation, where the generation of metadata is restricted to a limited number element due to the lackof training and professional knowledge, this paper points out a possible direction to resolve thisobstacle. The future metadata generation will thus be based on a top-down approach that can bothimprove the efficiency for generation and quality of contents.