diff --git a/vm/src/builtins.rs b/vm/src/builtins.rs
index 2614232e0..7bbaad581 100644
--- a/vm/src/builtins.rs
+++ b/vm/src/builtins.rs
@@ -408,13 +408,11 @@ fn builtin_len(vm: &mut VirtualMachine, args: PyFuncArgs) -> PyResult {
     let len_method_name = "__len__".to_string();
     match vm.get_method(obj.clone(), &len_method_name) {
         Ok(value) => vm.invoke(value, PyFuncArgs::default()),
-        Err(..) => Err(vm.context().new_str(
-            format!(
-                "TypeError: object of this {:?} type has no method {:?}",
-                obj, len_method_name
-            )
-            .to_string(),
-        )),
+        Err(..) => Err(vm.new_type_error(format!(
+            "object of type '{}' has no method {:?}",
+            objtype::get_type_name(&obj.typ()),
+            len_method_name
+        ))),
     }
 }
 
@@ -596,13 +594,10 @@ fn builtin_ord(vm: &mut VirtualMachine, args: PyFuncArgs) -> PyResult {
     let string = objstr::get_value(string);
     let string_len = string.chars().count();
     if string_len > 1 {
-        return Err(vm.new_type_error(
-            format!(
-                "ord() expected a character, but string of length {} found",
-                string_len
-            )
-            .to_string(),
-        ));
+        return Err(vm.new_type_error(format!(
+            "ord() expected a character, but string of length {} found",
+            string_len
+        )));
     }
     match string.chars().next() {
         Some(character) => Ok(vm
diff --git a/vm/src/frame.rs b/vm/src/frame.rs
index 19fa17bd7..89a8ccc3e 100644
--- a/vm/src/frame.rs
+++ b/vm/src/frame.rs
@@ -511,8 +511,8 @@ impl Frame {
                     Err(exception)
                 } else {
                     let msg = format!(
-                        "Can only raise BaseException derived types, not {:?}",
-                        exception
+                        "Can only raise BaseException derived types, not {}",
+                        exception.borrow()
                     );
                     let type_error_type = vm.ctx.exceptions.type_error.clone();
                     let type_error = vm.new_exception(type_error_type, msg);
@@ -850,7 +850,7 @@ impl Frame {
                 scope = scope.get_parent();
             } else {
                 let name_error_type = vm.ctx.exceptions.name_error.clone();
-                let msg = format!("Has not attribute '{}'", name);
+                let msg = format!("name '{}' is not defined", name);
                 let name_error = vm.new_exception(name_error_type, msg);
                 break Err(name_error);
             }
diff --git a/vm/src/obj/objcomplex.rs b/vm/src/obj/objcomplex.rs
index 59e13c772..ee119910c 100644
--- a/vm/src/obj/objcomplex.rs
+++ b/vm/src/obj/objcomplex.rs
@@ -68,7 +68,7 @@ fn complex_add(vm: &mut VirtualMachine, args: PyFuncArgs) -> PyResult {
     if objtype::isinstance(i2, &vm.ctx.complex_type()) {
         Ok(vm.ctx.new_complex(v1 + get_value(i2)))
     } else {
-        Err(vm.new_type_error(format!("Cannot add {:?} and {:?}", i, i2)))
+        Err(vm.new_type_error(format!("Cannot add {} and {}", i.borrow(), i2.borrow())))
     }
 }
 
diff --git a/vm/src/obj/objfloat.rs b/vm/src/obj/objfloat.rs
index edd495f38..39fd5d74d 100644
--- a/vm/src/obj/objfloat.rs
+++ b/vm/src/obj/objfloat.rs
@@ -52,7 +52,7 @@ pub fn make_float(vm: &mut VirtualMachine, obj: &PyObjectRef) -> Result<f64, PyO
         )?;
         Ok(get_value(&res))
     } else {
-        Err(vm.new_type_error(format!("Cannot cast {:?} to float", obj)))
+        Err(vm.new_type_error(format!("Cannot cast {} to float", obj.borrow())))
     }
 }
 
@@ -159,7 +159,7 @@ fn float_add(vm: &mut VirtualMachine, args: PyFuncArgs) -> PyResult {
             .ctx
             .new_float(v1 + objint::get_value(i2).to_f64().unwrap()))
     } else {
-        Err(vm.new_type_error(format!("Cannot add {:?} and {:?}", i, i2)))
+        Err(vm.new_type_error(format!("Cannot add {} and {}", i.borrow(), i2.borrow())))
     }
 }
 
@@ -176,7 +176,11 @@ fn float_divmod(vm: &mut VirtualMachine, args: PyFuncArgs) -> PyResult {
         let r2 = float_mod(vm, args.clone());
         Ok(vm.ctx.new_tuple(vec![r1.unwrap(), r2.unwrap()]))
     } else {
-        Err(vm.new_type_error(format!("Cannot divmod power {:?} and {:?}", i, i2)))
+        Err(vm.new_type_error(format!(
+            "Cannot divmod power {} and {}",
+            i.borrow(),
+            i2.borrow()
+        )))
     }
 }
 
@@ -193,7 +197,11 @@ fn float_floordiv(vm: &mut VirtualMachine, args: PyFuncArgs) -> PyResult {
             .ctx
             .new_float((get_value(i) / objint::get_value(i2).to_f64().unwrap()).floor()))
     } else {
-        Err(vm.new_type_error(format!("Cannot floordiv {:?} and {:?}", i, i2)))
+        Err(vm.new_type_error(format!(
+            "Cannot floordiv {} and {}",
+            i.borrow(),
+            i2.borrow()
+        )))
     }
 }
 
@@ -211,7 +219,7 @@ fn float_sub(vm: &mut VirtualMachine, args: PyFuncArgs) -> PyResult {
             .ctx
             .new_float(v1 - objint::get_value(i2).to_f64().unwrap()))
     } else {
-        Err(vm.new_type_error(format!("Cannot add {:?} and {:?}", i, i2)))
+        Err(vm.new_type_error(format!("Cannot add {} and {}", i.borrow(), i2.borrow())))
     }
 }
 
@@ -228,7 +236,7 @@ fn float_mod(vm: &mut VirtualMachine, args: PyFuncArgs) -> PyResult {
             .ctx
             .new_float(get_value(i) % objint::get_value(i2).to_f64().unwrap()))
     } else {
-        Err(vm.new_type_error(format!("Cannot mod {:?} and {:?}", i, i2)))
+        Err(vm.new_type_error(format!("Cannot mod {} and {}", i.borrow(), i2.borrow())))
     }
 }
 
@@ -254,7 +262,7 @@ fn float_pow(vm: &mut VirtualMachine, args: PyFuncArgs) -> PyResult {
         let result = v1.powf(objint::get_value(i2).to_f64().unwrap());
         Ok(vm.ctx.new_float(result))
     } else {
-        Err(vm.new_type_error(format!("Cannot add {:?} and {:?}", i, i2)))
+        Err(vm.new_type_error(format!("Cannot add {} and {}", i.borrow(), i2.borrow())))
     }
 }
 
diff --git a/vm/src/obj/objint.rs b/vm/src/obj/objint.rs
index 7d2470f6a..29e7e1515 100644
--- a/vm/src/obj/objint.rs
+++ b/vm/src/obj/objint.rs
@@ -188,9 +188,9 @@ fn int_add(vm: &mut VirtualMachine, args: PyFuncArgs) -> PyResult {
     arg_check!(
         vm,
         args,
-        required = [(i, Some(vm.ctx.int_type())), (i2, None)]
+        required = [(_i, Some(vm.ctx.int_type())), (i2, None)]
     );
-    let i = BigInt::from_pyobj(i);
+    let i = BigInt::from_pyobj(_i);
     if objtype::isinstance(i2, &vm.ctx.int_type()) {
         Ok(vm.ctx.new_int(i + get_value(i2)))
     } else if objtype::isinstance(i2, &vm.ctx.float_type()) {
@@ -198,7 +198,7 @@ fn int_add(vm: &mut VirtualMachine, args: PyFuncArgs) -> PyResult {
             .ctx
             .new_float(i.to_f64().unwrap() + objfloat::get_value(i2)))
     } else {
-        Err(vm.new_type_error(format!("Cannot add {:?} and {:?}", i, i2)))
+        Err(vm.new_type_error(format!("Cannot add {} and {}", _i.borrow(), i2.borrow())))
     }
 }
 
@@ -217,7 +217,11 @@ fn int_floordiv(vm: &mut VirtualMachine, args: PyFuncArgs) -> PyResult {
     if objtype::isinstance(i2, &vm.ctx.int_type()) {
         Ok(vm.ctx.new_int(get_value(i) / get_value(i2)))
     } else {
-        Err(vm.new_type_error(format!("Cannot floordiv {:?} and {:?}", i, i2)))
+        Err(vm.new_type_error(format!(
+            "Cannot floordiv {} and {}",
+            i.borrow(),
+            i2.borrow()
+        )))
     }
 }
 
@@ -243,9 +247,9 @@ fn int_sub(vm: &mut VirtualMachine, args: PyFuncArgs) -> PyResult {
     arg_check!(
         vm,
         args,
-        required = [(i, Some(vm.ctx.int_type())), (i2, None)]
+        required = [(_i, Some(vm.ctx.int_type())), (i2, None)]
     );
-    let i = BigInt::from_pyobj(i);
+    let i = BigInt::from_pyobj(_i);
     if objtype::isinstance(i2, &vm.ctx.int_type()) {
         Ok(vm.ctx.new_int(i - get_value(i2)))
     } else if objtype::isinstance(i2, &vm.ctx.float_type()) {
@@ -253,7 +257,11 @@ fn int_sub(vm: &mut VirtualMachine, args: PyFuncArgs) -> PyResult {
             .ctx
             .new_float(i.to_f64().unwrap() - objfloat::get_value(i2)))
     } else {
-        Err(vm.new_not_implemented_error(format!("Cannot substract {:?} and {:?}", i, i2)))
+        Err(vm.new_not_implemented_error(format!(
+            "Cannot substract {} and {}",
+            _i.borrow(),
+            i2.borrow()
+        )))
     }
 }
 
@@ -270,7 +278,11 @@ fn int_mul(vm: &mut VirtualMachine, args: PyFuncArgs) -> PyResult {
             .ctx
             .new_float(get_value(i).to_f64().unwrap() * objfloat::get_value(i2)))
     } else {
-        Err(vm.new_type_error(format!("Cannot multiply {:?} and {:?}", i, i2)))
+        Err(vm.new_type_error(format!(
+            "Cannot multiply {} and {}",
+            i.borrow(),
+            i2.borrow()
+        )))
     }
 }
 
@@ -290,7 +302,7 @@ fn int_truediv(vm: &mut VirtualMachine, args: PyFuncArgs) -> PyResult {
             .ctx
             .new_float(v1.to_f64().unwrap() / objfloat::get_value(i2)))
     } else {
-        Err(vm.new_type_error(format!("Cannot divide {:?} and {:?}", i, i2)))
+        Err(vm.new_type_error(format!("Cannot divide {} and {}", i.borrow(), i2.borrow())))
     }
 }
 
@@ -304,7 +316,7 @@ fn int_mod(vm: &mut VirtualMachine, args: PyFuncArgs) -> PyResult {
     if objtype::isinstance(i2, &vm.ctx.int_type()) {
         Ok(vm.ctx.new_int(v1 % get_value(i2)))
     } else {
-        Err(vm.new_type_error(format!("Cannot modulo {:?} and {:?}", i, i2)))
+        Err(vm.new_type_error(format!("Cannot modulo {} and {}", i.borrow(), i2.borrow())))
     }
 }
 
@@ -333,7 +345,11 @@ fn int_pow(vm: &mut VirtualMachine, args: PyFuncArgs) -> PyResult {
         let v2 = objfloat::get_value(i2);
         Ok(vm.ctx.new_float((v1.to_f64().unwrap()).powf(v2)))
     } else {
-        Err(vm.new_type_error(format!("Cannot raise power {:?} and {:?}", i, i2)))
+        Err(vm.new_type_error(format!(
+            "Cannot raise power {} and {}",
+            i.borrow(),
+            i2.borrow()
+        )))
     }
 }
 
@@ -349,7 +365,11 @@ fn int_divmod(vm: &mut VirtualMachine, args: PyFuncArgs) -> PyResult {
         let r2 = int_mod(vm, args.clone());
         Ok(vm.ctx.new_tuple(vec![r1.unwrap(), r2.unwrap()]))
     } else {
-        Err(vm.new_type_error(format!("Cannot divmod power {:?} and {:?}", i, i2)))
+        Err(vm.new_type_error(format!(
+            "Cannot divmod power {} and {}",
+            i.borrow(),
+            i2.borrow()
+        )))
     }
 }
 
@@ -364,7 +384,7 @@ fn int_xor(vm: &mut VirtualMachine, args: PyFuncArgs) -> PyResult {
         let v2 = get_value(i2);
         Ok(vm.ctx.new_int(v1 ^ v2))
     } else {
-        Err(vm.new_type_error(format!("Cannot xor {:?} and {:?}", i, i2)))
+        Err(vm.new_type_error(format!("Cannot xor {} and {}", i.borrow(), i2.borrow())))
     }
 }
 
@@ -379,7 +399,7 @@ fn int_or(vm: &mut VirtualMachine, args: PyFuncArgs) -> PyResult {
         let v2 = get_value(i2);
         Ok(vm.ctx.new_int(v1 | v2))
     } else {
-        Err(vm.new_type_error(format!("Cannot or {:?} and {:?}", i, i2)))
+        Err(vm.new_type_error(format!("Cannot or {} and {}", i.borrow(), i2.borrow())))
     }
 }
 
@@ -394,7 +414,7 @@ fn int_and(vm: &mut VirtualMachine, args: PyFuncArgs) -> PyResult {
         let v2 = get_value(i2);
         Ok(vm.ctx.new_int(v1 & v2))
     } else {
-        Err(vm.new_type_error(format!("Cannot and {:?} and {:?}", i, i2)))
+        Err(vm.new_type_error(format!("Cannot and {} and {}", i.borrow(), i2.borrow())))
     }
 }
 
diff --git a/vm/src/obj/objlist.rs b/vm/src/obj/objlist.rs
index e0a7b5e34..435542648 100644
--- a/vm/src/obj/objlist.rs
+++ b/vm/src/obj/objlist.rs
@@ -86,7 +86,7 @@ fn list_add(vm: &mut VirtualMachine, args: PyFuncArgs) -> PyResult {
         let elements = e1.iter().chain(e2.iter()).map(|e| e.clone()).collect();
         Ok(vm.ctx.new_list(elements))
     } else {
-        Err(vm.new_type_error(format!("Cannot add {:?} and {:?}", o, o2)))
+        Err(vm.new_type_error(format!("Cannot add {} and {}", o.borrow(), o2.borrow())))
     }
 }
 
diff --git a/vm/src/obj/objobject.rs b/vm/src/obj/objobject.rs
index 68a991e6a..f583f85a1 100644
--- a/vm/src/obj/objobject.rs
+++ b/vm/src/obj/objobject.rs
@@ -164,7 +164,7 @@ fn object_getattribute(vm: &mut VirtualMachine, args: PyFuncArgs) -> PyResult {
             let attribute_error = vm.context().exceptions.attribute_error.clone();
             Err(vm.new_exception(
                 attribute_error,
-                format!("{:?} object has no attribute {}", cls, name),
+                format!("{} has no attribute '{}'", obj.borrow(), name),
             ))
         }
     }
diff --git a/vm/src/obj/objset.rs b/vm/src/obj/objset.rs
index 2d9d4e6bd..05aebf4ed 100644
--- a/vm/src/obj/objset.rs
+++ b/vm/src/obj/objset.rs
@@ -58,7 +58,7 @@ fn set_new(vm: &mut VirtualMachine, args: PyFuncArgs) -> PyResult {
     );
 
     if !objtype::issubclass(cls, &vm.ctx.set_type()) {
-        return Err(vm.new_type_error(format!("{:?} is not a subtype of set", cls)));
+        return Err(vm.new_type_error(format!("{} is not a subtype of set", cls.borrow())));
     }
 
     let elements = match iterable {
diff --git a/vm/src/obj/objstr.rs b/vm/src/obj/objstr.rs
index 76e5c6538..e68d29a11 100644
--- a/vm/src/obj/objstr.rs
+++ b/vm/src/obj/objstr.rs
@@ -144,7 +144,7 @@ fn str_add(vm: &mut VirtualMachine, args: PyFuncArgs) -> PyResult {
             .ctx
             .new_str(format!("{}{}", get_value(&s), get_value(&s2))))
     } else {
-        Err(vm.new_type_error(format!("Cannot add {:?} and {:?}", s, s2)))
+        Err(vm.new_type_error(format!("Cannot add {} and {}", s.borrow(), s2.borrow())))
     }
 }
 
@@ -273,7 +273,11 @@ fn str_mul(vm: &mut VirtualMachine, args: PyFuncArgs) -> PyResult {
         }
         Ok(vm.ctx.new_str(result))
     } else {
-        Err(vm.new_type_error(format!("Cannot multiply {:?} and {:?}", s, s2)))
+        Err(vm.new_type_error(format!(
+            "Cannot multiply {} and {}",
+            s.borrow(),
+            s2.borrow()
+        )))
     }
 }
 
diff --git a/vm/src/obj/objtuple.rs b/vm/src/obj/objtuple.rs
index 541d9d276..d06cd1d2a 100644
--- a/vm/src/obj/objtuple.rs
+++ b/vm/src/obj/objtuple.rs
@@ -85,7 +85,7 @@ fn tuple_new(vm: &mut VirtualMachine, args: PyFuncArgs) -> PyResult {
     );
 
     if !objtype::issubclass(cls, &vm.ctx.tuple_type()) {
-        return Err(vm.new_type_error(format!("{:?} is not a subtype of tuple", cls)));
+        return Err(vm.new_type_error(format!("{} is not a subtype of tuple", cls.borrow())));
     }
 
     let elements = if let Some(iterable) = iterable {
diff --git a/vm/src/obj/objtype.rs b/vm/src/obj/objtype.rs
index bc62244ae..930e9621f 100644
--- a/vm/src/obj/objtype.rs
+++ b/vm/src/obj/objtype.rs
@@ -205,7 +205,7 @@ pub fn type_getattribute(vm: &mut VirtualMachine, args: PyFuncArgs) -> PyResult
             let attribute_error = vm.context().exceptions.attribute_error.clone();
             Err(vm.new_exception(
                 attribute_error,
-                format!("{:?} object {:?} has no attribute {}", mcl, cls, name),
+                format!("{} has no attribute '{}'", cls.borrow(), name),
             ))
         }
     }
diff --git a/vm/src/pyobject.rs b/vm/src/pyobject.rs
index 6e21c269e..a407e1409 100644
--- a/vm/src/pyobject.rs
+++ b/vm/src/pyobject.rs
@@ -25,7 +25,7 @@ use super::vm::VirtualMachine;
 use num_bigint::BigInt;
 use num_complex::Complex64;
 use num_traits::{One, Zero};
-use std::cell::{Ref, RefCell};
+use std::cell::RefCell;
 use std::collections::HashMap;
 use std::fmt;
 use std::rc::{Rc, Weak};
@@ -65,12 +65,25 @@ pub type PyObjectWeakRef = Weak<RefCell<PyObject>>;
 /// since exceptions are also python objects.
 pub type PyResult = Result<PyObjectRef, PyObjectRef>; // A valid value, or an exception
 
-/*
-impl fmt::Display for PyObjectRef {
+impl fmt::Display for PyObject {
     fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result {
-        write!(f, "Obj {:?}", self)
+        use self::TypeProtocol;
+        match &self.kind {
+            PyObjectKind::Module { name, .. } => write!(f, "module '{}'", name),
+            PyObjectKind::Class { name, dict,.. } => {
+                let type_name = objtype::get_type_name(&self.typ());
+                // We don't have access to a vm, so just assume that if its parent's name
+                // is type, it's a type
+                if type_name == "type" {
+                    write!(f, "type object '{}'", name)
+                } else {
+                    write!(f, "'{}' object", type_name)
+                }
+            }
+            _ => write!(f, "'{}' object", objtype::get_type_name(&self.typ())),
+        }
     }
-}*/
+}
 
 /*
  // Idea: implement the iterator trait upon PyObjectRef
@@ -588,7 +601,13 @@ pub trait TypeProtocol {
 
 impl TypeProtocol for PyObjectRef {
     fn typ(&self) -> PyObjectRef {
-        match self.borrow().typ {
+        self.borrow().typ()
+    }
+}
+
+impl TypeProtocol for PyObject {
+    fn typ(&self) -> PyObjectRef {
+        match self.typ {
             Some(ref typ) => typ.clone(),
             None => panic!("Object {:?} doesn't have a type!", self),
         }
diff --git a/vm/src/vm.rs b/vm/src/vm.rs
index c2c7b561e..571ac42cc 100644
--- a/vm/src/vm.rs
+++ b/vm/src/vm.rs
@@ -464,10 +464,11 @@ impl VirtualMachine {
         let cls = obj.typ();
         match cls.get_attr(method_name) {
             Some(method) => self.call_get_descriptor(method, obj.clone()),
-            None => {
-                Err(self
-                    .new_type_error(format!("{:?} object has no method {:?}", obj, method_name)))
-            }
+            None => Err(self.new_type_error(format!(
+                "{} has no method {:?}",
+                obj.borrow(),
+                method_name
+            ))),
         }
     }