Guidelines on structuring large racket project

Question

Is there any general guidelines (equivalent to How to Write Go Code ) on how to structure large racket project ?

Answer 1

Your question is about structuring large Racket projects, however your example link How to Write Go Code seems mainly to be about how to create a package. These aren't necessarily the same thing (although you might divide a large Racket program into different source directories, which could be "collections" or "packages"). Anyway, to address both parts:

1. How to structure a large Racket project: Asumu's answer provides a good link: How to Program Racket. In particular see section 3, "Units of Code", which discusses modules. Also, one technique you will see in the source code of Racket itself, is that a module may exist solely to require functions from others and provide them -- to "re-provide". In other words you can use modules like this to chunk up other modules and selectively expose them as a "layer" in your structure. Furthermore, Racket has a class system if that fits your problem domain, and generics if you have some sort of "interface" or "protocol" strategy. And more. Really, there a many, many techniques available in Racket to structure a large project.

2. How to make Racket packages: This is an interesting time to ask. Historically Racket has used something called Planet as a package manager. More recently, it has a new package system. Although not yet officially out of beta, many people are already using it for real work. The current documentation for that, although in a different style than the Go doc, is Package Management in Racket Beta).

You asked for general guidelines, which is a bit open-ended. If you have any specific choices you're weighing for how to structure a large project, perhaps you could ask about those one by one to get crisper answers?