Improving API documentation using API usage information

Abstract

Abstract

What makes a good code example?: A study of programming Q&A in StackOverflow

“…The extracted code by all of these tools does not necessarily have a familiar context and is not always accompanied with helpful explanation. To overcome some of these shortcomings, Stylos et al developed a tool that extracts some common usage scenarios using web search (such as object creation) and injects them into the API documentation [26]; albeit for very limited scenario types.…”

Recommending reference API documentation

“…In contrast to this work, which focused on determining the quality of API reference documentation, we focus on the automated extraction of information in the documentation that can be effectively recommended during a task. Stylos et al (2009) proposed Jadeite, which studies source code and statistically provides recommendations to the programmers on the most used classes, constructors, methods, and objects. These specifications helped detect bugs which were introduced due to developers using APIs for purposes that were not intended by the API.…”

Using Association Metrics to Help Users Navigate API Documentation

“…The Jadeite Java documentation system gives programmers more cues about which classes and methods are commonly used [16]. Jadeite uses font sizes similar to tag clouds but in a single alphabetical list.…”

“…Popularity information for each item is computed with the same technique used in the Jadeite 1 http://mootools.net documentation tool [16]: The popularity for each package, class and method is based on the number of hits returned by a Google search for each item. For example, the weight for the InputStream class is computed based on the number of results returned for the query: "java.io" +InputStream.…”