This page provides some supprimental materials for the following paper.

Yoshiki Higo, and Shinji Kusumoto, "How Should We Measure Functional Sameness from Program Source Code? -- An Exploratory Study on Java Methods --", submitted to FSE2014.

We used two datasets in the experiment of the above paper. Those datasets are accessible: APACHE and UCI

We also put some figures in this page because in the paper we were able to put figures for APACHE dataset due to space limitation

Figure 2: Overview of Three Types of Similarities for APACHE and UCI (Each dot represents a method pair. Each black dot is a method pair whose names are different and each red dot is a method pair whose names are the same)

APACHE

UCI

Figure 3(a): Boxplot representing relationship structural similarity distribution and name abstractness. X-axis represents the degree of name abstractness. There are

APACHE	UCI

Figure 3(b): Boxplot representing relationship vocabulary similarity distribution and name abstractness. X-axis represents the degree of name abstractness. There are

APACHE	UCI

Figure 5: Vocabulary similarity in each category

APACHE	UCI

Figure 6: Structural-similarity-based and vocabulary-similarity-based boxplots

APACHE

UCI