Automatically Finding Web Documents Related to a Code Sample

概要

論文の詳細を見る
Code search engines on Web have played important roles for novice programmers during developing software or studying programming by reference examples. However, they have faced a problem from search results because these code samples lack context or can be unfamiliar code snippets that are incomprehensible. They must find external information resources by using another Web search engine with numerous queries manually and too many pages can be visited. This paper presents a method to automatically find related Web documents to support them to comprehend source code. We utilize Web search service APIs for collecting related information of Web pages to process. Advantaged techniques of information retrieval are used to rank the text documents. For demonstrating benefits of this method, the proposed method is applied into an architecture model of a source code search engine and its implementation. Experimental results from several case studies showed that the method is useful in supporting programmers understanding new concepts in code samples.
一般社団法人情報処理学会の論文
2010-03-11