Mining Billions of AST Nodes to Study Actual and Potential Usage of Java Language Features

by Robert Dyer, Hridesh Rajan, Hoan Anh Nguyen, and Tien N. Nguyen

Table of Contents

• Introduction
• Section 2 - JLS Background
• Section 4.2 - Dataset Metrics
• Section 5.1 - RQ1 - Do projects use new language features before the features are released?
• Section 5.2 - RQ2 - How frequently is each language feature used?
  • Section 5.2.1 - Investigating Frequently Used Features
• Section 5.3 - RQ3 - How did committers adopt and use language features?
  • Section 5.3.1 - RQ3.1 - How many committers adopted and used new features over time?
  • Section 5.3.2 - RQ3.2 - How much did committers use each feature?
  • Section 5.3.3 - RQ3.3 - Did committers adopt features on an individual basis or as a team?
  • Section 5.3.4 - RQ3.4 - Did committers use all new features?
• Section 5.4 - RQ4 - Were there missed opportunities to use language features?
• Section 5.5 - RQ5 - Was old code converted to use new language features?

Section 5.4 - RQ4 - Were there missed opportunities to use language features?

Boa Program: potential-uses.boa
Raw Data: potential-uses.txt
Figure 15 Generator: potentialuses.java

This material mines potential uses for language features, shown in Figure 15.

The output contains five keys (each key represents a row in Figure 15):

Key	Description
OldPotential	Potential uses in old files.
NewPotential	Potential uses in new files.
AllPotential	Potential uses in all files.
FilesPotential	Number of files with potential uses.
ProjectsPotential	Number of projects with potential uses.

and each key contains an index for the feature being mined:

Key Index	Column in Figure 6
Assert	Assert
Varargs	Varargs
BinaryLit	Binary Literals
Diamond	Diamond
MultiCatch	MultiCatch
Underscore	Underscore Literals

Boa Program: try-resources-bugs.boa
Raw Data: try-resources-bugs.txt

This material mines potential bugs from not using try with resources.

Last modified: $Id: index.html,v 1.2 2013/09/20 21:49:25 rdyer Exp $