Motivated by a need for a model of reversible computation appropriate for a Brownian molecular architecture, the $\aleph$ calculus is introduced. This novel model is declarative, concurrent, and term-based--encapsulating all information about the program data and state within a single structure in order to obviate the need for a von Neumann-style discrete computational 'machine', a challenge in a molecular environment. The name is inspired by the Greek for 'not forgotten', due to the emphasis on (reversibly) learning and un-learning knowledge of different variables. To demonstrate its utility for this purpose, as well as its elegance as a programming language, a number of examples are presented; two of these examples, addition/subtraction and squaring/square-rooting, are furnished with designs for abstract molecular implementations. A natural by-product of these examples and accompanying syntactic sugar is the design of a fully-fledged programming language, alethe, which is also presented along with an interpreter. Efficiently simulating $\aleph$ on a deterministic computer necessitates some static analysis of programs within the alethe interpreter in order to render the declarative programs sequential. Finally, work towards a type system appropriate for such a reversible, declarative model of computation is presented.
翻译:由于需要一种适合布朗分子结构的可逆计算模型,因此引入了$=aleph$计算模型。这个新颖模型是宣示性的、同时的和基于术语的,在单一结构内包含关于程序数据和状态的所有信息,以避免需要冯纽曼式的离散计算“机器”,这是分子环境中的一个挑战。这个名称来自希腊人“不被遗忘”的灵感,因为强调(不可逆转的)学习和不学习不同变量的知识。为了证明它在这方面的实用性,以及它作为一种编程语言的优雅性,我们举出了一些例子;其中两个例子,即添加/增缩和方略/方根,都配有抽象分子执行的设计。这些例子的自然副产品和伴随的合成糖是设计一种完全成熟的编程语言,Alethe(也与译员一起展示)。在确定性计算机的系统类型解释程序中,在确定性格式化的顺序上高效地模拟$\leph$,最后将一个排序的顺序转化为一个适当的递定式的计算机结构分析。