diff --git a/ChangeLog b/ChangeLog
index 1d1968862..4b35164af 100644
--- a/ChangeLog
+++ b/ChangeLog
@@ -1,3 +1,16 @@
+2013-01-12  Alexei Podtelezhnikov  <apodtele@gmail.com>
+
+	[truetype] Kill very old vector normalization hacks.
+
+	Back in the days, vector length calculations were not very accurate
+	and the vector normalization function, Normalize, had to meticulously
+	correct the errors for long vectors [commit b7ef2b096867]. It was no
+	longer necessary after accurate CORDIC algorithms were adopted, but
+	the code remained. It is time to kill it.
+
+	* src/truetype/ttinterp.c (Normalize): Remove error compensation.
+	(TT_VecLen): Remove any mention of old less accurate implementation.
+
 2013-01-11  Werner Lemberg  <wl@gnu.org>
 
 	Disable FT_CONFIG_OPTION_OLD_INTERNALS.
diff --git a/src/truetype/ttinterp.c b/src/truetype/ttinterp.c
index 19fb66c17..f86232080 100644
--- a/src/truetype/ttinterp.c
+++ b/src/truetype/ttinterp.c
@@ -1543,79 +1543,6 @@
 
 
   /* return length of given vector */
-
-#if 0
-
-  static FT_Int32
-  TT_VecLen( FT_Int32  x,
-             FT_Int32  y )
-  {
-    FT_Int32   m, hi1, hi2, hi;
-    FT_UInt32  l, lo1, lo2, lo;
-
-
-    /* compute x*x as 64-bit value */
-    lo = (FT_UInt32)( x & 0xFFFFU );
-    hi = x >> 16;
-
-    l  = lo * lo;
-    m  = hi * lo;
-    hi = hi * hi;
-
-    lo1 = l + (FT_UInt32)( m << 17 );
-    hi1 = hi + ( m >> 15 ) + ( lo1 < l );
-
-    /* compute y*y as 64-bit value */
-    lo = (FT_UInt32)( y & 0xFFFFU );
-    hi = y >> 16;
-
-    l  = lo * lo;
-    m  = hi * lo;
-    hi = hi * hi;
-
-    lo2 = l + (FT_UInt32)( m << 17 );
-    hi2 = hi + ( m >> 15 ) + ( lo2 < l );
-
-    /* add them to get 'x*x+y*y' as 64-bit value */
-    lo = lo1 + lo2;
-    hi = hi1 + hi2 + ( lo < lo1 );
-
-    /* compute the square root of this value */
-    {
-      FT_UInt32  root, rem, test_div;
-      FT_Int     count;
-
-
-      root = 0;
-
-      {
-        rem   = 0;
-        count = 32;
-        do
-        {
-          rem      = ( rem << 2 ) | ( (FT_UInt32)hi >> 30 );
-          hi       = (  hi << 2 ) | (            lo >> 30 );
-          lo     <<= 2;
-          root   <<= 1;
-          test_div = ( root << 1 ) + 1;
-
-          if ( rem >= test_div )
-          {
-            rem  -= test_div;
-            root += 1;
-          }
-        } while ( --count );
-      }
-
-      return (FT_Int32)root;
-    }
-  }
-
-#else
-
-  /* this version uses FT_Vector_Length which computes the same value */
-  /* much, much faster..                                              */
-  /*                                                                  */
   static FT_F26Dot6
   TT_VecLen( FT_F26Dot6  X,
              FT_F26Dot6  Y )
@@ -1629,8 +1556,6 @@
     return FT_Vector_Length( &v );
   }
 
-#endif
-
 
   /*************************************************************************/
   /*                                                                       */
@@ -2718,7 +2643,6 @@
                       FT_UnitVector*  R )
   {
     FT_F26Dot6  W;
-    FT_Bool     S1, S2;
 
     FT_UNUSED_EXEC;
 
@@ -2728,80 +2652,18 @@
       Vx *= 0x100;
       Vy *= 0x100;
 
-      W = TT_VecLen( Vx, Vy );
-
-      if ( W == 0 )
+      if ( Vx == 0 && Vy == 0 )
       {
         /* XXX: UNDOCUMENTED! It seems that it is possible to try   */
         /*      to normalize the vector (0,0).  Return immediately. */
         return SUCCESS;
       }
-
-      R->x = (FT_F2Dot14)TT_DivFix14( Vx, W );
-      R->y = (FT_F2Dot14)TT_DivFix14( Vy, W );
-
-      return SUCCESS;
     }
 
     W = TT_VecLen( Vx, Vy );
 
-    Vx = TT_DivFix14( Vx, W );
-    Vy = TT_DivFix14( Vy, W );
-
-    W = Vx * Vx + Vy * Vy;
-
-    /* Now, we want that Sqrt( W ) = 0x4000 */
-    /* Or 0x10000000 <= W < 0x10004000      */
-
-    if ( Vx < 0 )
-    {
-      Vx = -Vx;
-      S1 = TRUE;
-    }
-    else
-      S1 = FALSE;
-
-    if ( Vy < 0 )
-    {
-      Vy = -Vy;
-      S2 = TRUE;
-    }
-    else
-      S2 = FALSE;
-
-    while ( W < 0x10000000L )
-    {
-      /* We need to increase W by a minimal amount */
-      if ( Vx < Vy )
-        Vx++;
-      else
-        Vy++;
-
-      W = Vx * Vx + Vy * Vy;
-    }
-
-    while ( W >= 0x10004000L )
-    {
-      /* We need to decrease W by a minimal amount */
-      if ( Vx < Vy )
-        Vx--;
-      else
-        Vy--;
-
-      W = Vx * Vx + Vy * Vy;
-    }
-
-    /* Note that in various cases, we can only  */
-    /* compute a Sqrt(W) of 0x3FFF, eg. Vx = Vy */
-
-    if ( S1 )
-      Vx = -Vx;
-
-    if ( S2 )
-      Vy = -Vy;
-
-    R->x = (FT_F2Dot14)Vx;   /* Type conversion */
-    R->y = (FT_F2Dot14)Vy;   /* Type conversion */
+    R->x = (FT_F2Dot14)TT_DivFix14( Vx, W );
+    R->y = (FT_F2Dot14)TT_DivFix14( Vy, W );
 
     return SUCCESS;
   }