diff --git a/.gitignore b/.gitignore
index 3aa26ca932ec7ab6ea370d0af281abeb5f60dafe..7e4e4f0c2ef39f99a65bc9eb8d7fba1da5953b9b 100644
--- a/.gitignore
+++ b/.gitignore
@@ -9,6 +9,6 @@
*~
*.bak
.coq-native/
-Makefile.coq
+Makefile.coq*
*.crashcoqide
-html/
\ No newline at end of file
+html/
diff --git a/descr b/descr
new file mode 100644
index 0000000000000000000000000000000000000000..fa9d5c7bd4508343fad1febc1f520feb2eb9aa5a
--- /dev/null
+++ b/descr
@@ -0,0 +1,19 @@
+This project contains an extended "Standard Library" for Coq called coq-std++.
+The key features of this library are as follows:
+
+- It provides a great number of definitions and lemmas for common data
+ structures such as lists, finite maps, finite sets, and finite multisets.
+- It uses type classes for common notations (like `∅`, `∪`, and Haskell-style
+ monad notations) so that these can be overloaded for different data structures.
+- It uses type classes to keep track of common properties of types, like it
+ having decidable equality or being countable or finite.
+- Most data structures are represented in canonical ways so that Leibniz
+ equality can be used as much as possible (for example, for maps we have
+ `m1 = m2` iff `∀ i, m1 !! i = m2 !! i`). On top of that, the library provides
+ setoid instances for most types and operations.
+- It provides various tactics for common tasks, like an ssreflect inspired
+ `done` tactic for finishing trivial goals, a simple breadth-first solver
+ `naive_solver`, an equality simplifier `simplify_eq`, a solver `solve_proper`
+ for proving compatibility of functions with respect to relations, and a solver
+ `set_solver` for goals involving set operations.
+- It is entirely dependency- and axiom-free.