International Credit Insurance & Surety Association (ICISA)
As more new languages are evolved and libraries developed, it has become
increasingly difficult for programmers to get familiar with all the usage of them. Therefore,
it will be of great help if the programmers can find the desired functions or even better the
sample source codes as soon as they want. While there are already many tools developed for
document retrieval, most of them are only suitable for text documents, not for source codes.
To accommodate this increasing demand for source code retrieval for programmers, we
propose a search system for sample programs. It is based on a revised Bayesian belief
network model which allows the programmer to use free text in either Chinese or English to
locate the right function and samples that they want. A comparison with traditional search
models is made.