Fix documentation typo

python/dune/perftool/sumfact/sumfact.py

@@ -364,8 +364,8 @@ def sum_factorization_kernel(a_matrices,
     A_l * X --> [m_l,n_l] * [n_l, ...] = [m_l,n_{l+1},...,n_{d-1},m_0,...,m_{l-1}]
     R_l (A_l*X) --> [n_{l+1},...,n_{d-1},m_0,...,m_{l-1}]
 
-    So the multiplication with A_l is reduction over one index and the
-    transformation brings the next reduction index in the fastest
+    So the multiplication with A_l is a reduction over one index and
+    the transformation brings the next reduction index in the fastest
     position.
 
     Note: In the code below the transformation step is directly done