[dcl.array] No longer explain array subscript in terms of array-to-pointer conversion

source/declarations.tex

@@ -3452,25 +3452,17 @@
 The expression
 \tcode{x3d[i]}
 is equivalent to
-\tcode{*(x3d + i)};
-in that expression,
-\tcode{x3d}
-is subject to the array-to-pointer conversion\iref{conv.array}
-and is first converted to
-a pointer to a 2-dimensional
-array with rank
-$5 \times 7$
-that points to the first element of \tcode{x3d}.
-Then \tcode{i} is added,
-which on typical implementations involves multiplying
+\tcode{*(x3d + i)},
+and is an lvalue denoting
+the $\tcode{i}^\text{th}$ array element of
+\tcode{x3d},
+which is an array with rank $5 \times 7$.
+On typical implementations,
+computing the address of this two-dimensional array
+involves multiplying
 \tcode{i} by the
-length of the object to which the pointer points,
+length of the array in bytes,
 which is \tcode{sizeof(int)}$ \times 5 \times 7$.
-The result of the addition and indirection is
-an lvalue denoting
-the $\tcode{i}^\text{th}$ array element of
-\tcode{x3d}
-(an array of five arrays of seven integers).
 If there is another subscript,
 the same argument applies again, so
 \tcode{x3d[i][j]} is