Correspondent computing
Pen-Nan Lee · 1988
In the context of the preliminary study reported here, correspondent computing is loosely defined as the generation of effects (results or outcomes) which are correspondent to the effect of an operation in a computer program. The operation whose effect is the target of correspondent computing is called the primary operation. Operations generate correspondent effects are called correspondent operations. The philosophy of correspondence is that if executing an operation can produce a significant effect, an equivalently significant effect may be generated by another semantically correspondent operation. In general, correspondent operations can be categorized as the following three types. 1) Reciprocal type. The behaviour of a reciprocal operation is semantically the inverse of the primary operation. 2) Duplicate type. The behaviour of a duplicate operation is semantically equivalent to the primary operation. 3) Residual (or complementary) type. An operation that fulfills the definition of correspondent operations, but is neither a duplicate, nor a reciprocal is termed as residual operation. All these three types of correspondent operations can be easily derived from the primary operation on the basis of correspondence. We believe that Correspondent computing has the potential of becoming a powerful tool and can be used in the areas such as Fault Tolerance, Artificial Intelligence, Graphics, Database systems, etc. In the following paragraphs, we discuss the use of correspondent computing for Error Detection.