Usability challenges for enterprise service-oriented architecture APIs

Using Association Metrics to Help Users Navigate API Documentation

“…As described in Sections I and II, a previous study [2] indicates that programmers often have in mind a particular operation or action but do not know which classes or packages contain these methods. However, methods that achieve similar tasks in two different APIs are likely to have different names, and this vocabulary problem is a significant barrier to switching between APIs efficiently.…”

“…The program takes as inputs the implementation of the API (with Javadoc-style commenting) and a corpus of existing source code that uses that API. The Javadoc tool 2 For the association strength between two methods, we count the number of times they are called within seven method calls of each other. To choose this number, we analyzed how it affected the top associations for each method.…”

“…After twenty years of Java development, static Javadoc entries are still the dominant form of API references. Although modern programmers often use Google to search for API usage assistance, evidence indicates that this is an imperfect strategy [15] and is not an option for developers using proprietary or less-popular platforms [2].…”

Automatically Detecting Opportunities for Web Service Descriptions Improvement

“…Although the importance of well-designed WSDL document has been identified as a central concern in Web Service reuse [14,15], different studies have pointed out the existence of widespread problems in documentation [16], naming [11] and interface design [9,17] in real-life WSDL documents.…”

A case study of using HCI methods to improve tools for programmers

“…The Natural Programming project [1] at Carnegie Mellon University (CMU) has repeatedly demonstrated that many of the these techniques can be adapted to improve the effectiveness of the tools and languages used by developers. For example, we have applied them to the user interfaces of integrated development environments (IDEs) [2,3], debuggers [4], programming languages [5], and APIs (application programming interfaces) [6][7][8][9][10]. Many HCI techniques are applicable, and we have used contextual inquiry, heuristic evaluation, cognitive walkthrough, ethnography, and controlled lab studies.…”