Fixing what was presumably a typo in the naming conventions file.

Indeed, "forall x, op x x = x" in not in the list, while this is one
of the two standard meanings of idempotence. So, knowing that x, y,
... and not n are used elsewhere for variables names, and elt for
constants. Moreover, it is probable that before using consistently x,
y and z, I had also used m and n, sometimes. So, a convergent
probability that it is (just) a typo.

dev/doc/naming-conventions.tex

@@ -267,7 +267,7 @@ \subsection{Specific conventions}
 {forall x y:D, op (op' x y) = op' x (op y)}
 
 \itemrule{Idempotency of binary operator {\op} on domain {\D}}{Dop\_idempotent}
-{forall x:D, op x n = x}
+{forall x:D, op x x = x}
 
 \itemrule{Idempotency of unary operator {\op} on domain {\D}}{Dop\_idempotent}
 {forall x:D, op (op x) = op x}