| | | | Abstract: Collections of already developed programs are important resources for efficient development of reliable software systems. In this research, we propose a novel model of ranking software components, called component rank, based on analyzing actual use relations among the components and propagating the significance through the use relations. We have developed a Java class search system named SPARS-J using the component rank, and applied SPARS-J to various collections of Java programs. The result shows that SPARS-J ranks generic and more specific components higher. SPARS-J has also been applied to several companies, and it showed its capability of searching components in the company assets. | | | |