We present the concept of a functional programming language called VML (View Modeling Language), providing facilities to increase the efficiency of the iterative, trial-and-error cycle which frequently appears in any knowledge discovery process. In VML, functions can be specified so that returning values implicitly “remember”, with a special internal representation, that it was calculated from the corresponding function. VML also provides facilities for “matching” the remembered representation so that one can easily obtain, from a given value, the functions and/or parameters used to create the value. Further, we describe, as VML programs, successful knowledge discovery tasks which we have actually experienced in the biological domain, and argue that computational knowledge discovery experiments can be efficiently developed and conducted using this language.