Roll Your Own Web Database: An Innovative Approach for Providing Searchable Web Content

網際網路是一個巨大的資料庫，與常見的關聯式資料庫類比，要從今天的網際網路取得資訊，現有的搜尋方式並不充分。網頁資料相當於原始資料表，但關聯式資料庫具有View、SQL語法等等工具，網際網路幾乎只有關鍵字搜尋一項。HTML最初的設計理念，是針對資料的排版呈現，提供的是呈現層，而非資料層。本研究提出了Object-Oriented Schema Model（OOSM），這是一個文法模型。OOSM可由領域專家或一般的使用者來制定。使用者利用OOSM來標記網頁內容，OOSM核心可以將一般呈現用的網頁轉譯成可供資料處理的內容。本研究除了文法模型，亦實作了一個供使用者建立對應的工具，以便於使用。

The paper is aimed at addressing two issues: first, despite of the importance of semantic information in HTML pages, it is often ignored by search engines due to various technology difficulties; second, the ambiguity problem sometimes makes results returned by search engines much less useful. OOSM, a schema model as well as a set of information processing tools, is proposed in the paper. OOSM develops the concept of coarse mapping, that is, users are allowed (but not restricted) to associate a grammar node to a sub section instead of a single node on a HTML page. AS a result, minor modifications of the annotated HTML page can be tolerated. We believe that OOSM is a right solution for the issues presented in this paper.