A programming language interpreter system based on equivalent transformation

Both correctness and efficiency of software are becoming increasingly critical as computer systems become more important in our lives. Due to the emergence of ubiquitous computing, there are far more software systems than before and it is getting increasingly difficult to fix and tune all of them, according to their current situations, by hand. One solution to this problem is the use of the adaptive software which modifies itself to fit its environment. However, it seems to have no rigorous theory for correctness. In many cases, correct computations are regarded as a sequence of equivalent transformations. In the equivalent transformation (ET) computation model, all the procedures are represented by a sequence of equivalent transformations, and we can achieve both correctness and efficiency of programs by rigorous theory. We have developed ETL, which is a programming language to describe programs based on the model, and a programming language interpreter system ETI. In this paper, we present the ETI which can correctly and efficiently solve problems which are difficult to solve by conventional approaches.