My attempt to better understand both Coq and Category Theory by formalizing the latter in the former.
Switch branches/tags
Nothing to show
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
CT
scripts
.gitignore
.travis.yml
Makefile
README.md
_CoqProject
ct_ghpages_id_rsa.enc

README.md

ct Build Status

An attempt at a category theory encoding in Coq, for learning purposes.

Layout

Core concepts (Category, Functor, NaturalTransformation, etc.) are in the CT directory. Instances and derivations both live in the CT/Instance directory (and thus the CT.Instance namespace). A simple algebra hierarchy is found in CT/Algebra. Definitions there are largely decategorified, even when the structures can be defined purely categorically. There are instances (e.g. CT.Instance.Category.MonoidCategory (the category for any given Monoid)) which make use of them.

The layout of the project is fairly simple. If you wish to work with, for example, functor categories, you'll likely want to do something like:

Require Import CT.Category.
Require Import CT.Instance.Category.FunctorCategory.
(* ... *)

In any documentation we write, an instance is a particular usage of a concept. For example CT.Instance.Coq.Functor exports a CoqOptionFunctor which, as the name suggests, is a Functor for Coq's option type (specifically it is an endofunctor from CoqType -> CoqType, where CoqType is defined in CT.Instance.Coq.Category to be the category of Coq types).

We call something a derivation if passing it something generates a specialized version of that thing. For example, FunctorCategory is a derivation because it takes two Category parameters to return the functor category between them. Similarly, a MonoidCategory takes a Monoid and returns a Category with a single object (with composition defined by Monoid's mu).

These ("instance", "derivation") are mostly made-up terms, but we have been using them pretty consistently in our documentation/on IRC.

Notes on notation

We rarely use Notation or Infix in the library. This is intentional and done for added clarity. While Coq's notation system can be extremely helpful (especially in combination with scopes), since this is a library written for educational purposes, we wish to not cloud any view of what is actually happening. We do sprinkle in use of implicit arguments fairly liberally, however.

References

... so far.

License

MIT