Fix some documentation examples involving the repr of a float.
diff --git a/Doc/faq/design.rst b/Doc/faq/design.rst
index 94afaff..02e3fad 100644
--- a/Doc/faq/design.rst
+++ b/Doc/faq/design.rst
@@ -75,9 +75,9 @@
 function prints fewer digits and this often results in the more sensible number
 that was probably intended::
 
-   >>> 0.2
-   0.20000000000000001
-   >>> print 0.2
+   >>> 1.1 - 0.9
+   0.20000000000000007
+   >>> print 1.1 - 0.9
    0.2
 
 One of the consequences of this is that it is error-prone to compare the result