Mining Software Library Specifications using Inductive Logic Programming.



Sriram Sankaranarayanan, Franjo Ivancic, and Aarti Gupta
Software libraries are ubiquitous for organizing widely useful functionalities in order to promote modularity and code reuse. A typical software library is used by client programs through an API that hides library internals from the client. Typically, the rules governing the correct usage of the API are documented informally. As a result, the behaviour of the library under some corner cases may not be well understood by the programmer.

We propose a methodology for learning interface specifications using inductive logic programming (ILP). Our technique runs several unit tests on the library in order to generate relations describing the operation of the library. The data collected from these tests are used by an inductive learner to obtain rich Prolog specifications. Such specifications capture essential properties of the library. They may be used for applications such as reverse engineering, and also to construct checks on that enforce proper API usage.

ps    pdf
Intl. Conference On Software Engineering (ICSE 2008), pages 131-140, ACM Press.
Copyright (C) Association For Computer Machinery (ACM). See first page of paper for copyright notice. Copy has been made available online for personal use only. Do not redistribute without permission.


Sriram Sankaranarayanan