作者: Chris Goad
DOI: 10.1007/BFB0000060
关键词: Inductive programming 、 Programming paradigm 、 Computer science 、 Symbolic programming 、 Functional reactive programming 、 Computer programming 、 Reactive programming 、 Procedural programming 、 Theoretical computer science 、 Automatic programming
摘要: Abstract : According to the usual formulation of automatic programming task, one starts with a specification problem, and seeks automatically construct program satisfying that specification. This paper concerns different style programming. Rather than defining class problems be dealt by language in which those are formulated, we instead consider classes defined ordinary mathematical terms. Also, our aims from traditional interested primarily increasing efficiency computations, rather transferring burden human computer. Let alpha(p, x, y) ternary predicate. Suppose course some large computation obliged repeatedly compute values y given p x. further sequence p's x's treated, changes slowly x rapidly. Then seek synthesize fast special purpose A sub for each p; is expected when as input. We present example detail, namely, method synthesizing programs eliminating hidden surfaces displays three dimensional scenes. (Hidden surface elimination central computer graphics). In test method, synthetic specialized treating particular scene - but an arbitrary point view proved order magnitude faster best available general algorithm. (Author)