There are many technical procedures to analyze and evaluate raw source code of any software. However, most of them focus on productivity and efficiency of the code rather than the concealed historical value, especially with the legacy software and programs. On the other hand, preservation of digital and technological collection has increasingly been one of the major concerns in certain places like museum and science center. Such activity, therefore, includes the exploration of digital collection and extraction of the undiscovered or disappearing knowledge. This study displays an effort to experimentally use topic modeling approaches to extract interesting facts and knowledge within a special kind of digital-born collection, software’s source code. By exercising multiple techniques with selected subjects, this study also estimates the possibility to create a primitive framework to investigate digital collection, such as codebase, for further use in museum context.