Modules are linked to the existing "modules" feature of Mediawiki, written in Lua via Scribunto; it seems too limitative because: this project is independent of the language and does not directly describe the "implementation", but is only a repository of "interfaces" to describe these "functions"; the "functions" are designed to be pure, so they do not distinguish really *input* and *output * parameters but should allow forward and backward infererence in any direction (using multiple implementations in various languages, for each type of input/ouput, under the generic concept of "free variables" which may become "bound variables" independently to solve a problem; we are near from the concepts of IA and related languages working at this level, like Prolog, and as well the type of binding is not restricted to a single value but could be a set of values with probabilities; similar to the current researches on quantum physics, working with probabilities, intrication of states, undetermined/infinite number of internal states); it is also tied to the problem (and paradox) of compleness and algorithms complexity (think about Gödel's theorem on incompletude). For this project to be useful and productive, we'll need to go beynf a simple repository of code (implementations) and the imperfect vision of functions, which are a subclass of mathematical objects, and tied as well to the modelization of infinities and orders of magnitudes (or meta-orders). The purpose of this project is to help describe how we can *compose* the "functions" in a suitable representation as a graph to create larger objects, rather than describe how each component ("function") is implemented: the project could even (and should even be able to) describe functions that (still, for now) have NO implementation at all: very useful to manage projects with desired but still unmet goals. verdy_p (talk) 12:41, 29 September 2020 (UTC)[reply]