diff --git a/doc/assert.txt b/doc/assert.txt
index 7d1c21da3..191cd1a66 100644
--- a/doc/assert.txt
+++ b/doc/assert.txt
@@ -39,23 +39,6 @@ assertion fails you will see the value of ``x``::
     test_assert1.py:5: AssertionError
     ========================= 1 failed in 0.02 seconds =========================
 
-Reporting details about the failing assertion is achieved by re-evaluating
-the assert expression and recording the intermediate values.
-
-Note: If evaluating the assert expression has side effects you may get a
-warning that the intermediate values could not be determined safely.  A
-common example of this issue is an assertion which reads from a file::
-
-        assert f.read() != '...'
-
-If this assertion fails then the re-evaluation will probably succeed!
-This is because ``f.read()`` will return an empty string when it is
-called the second time during the re-evaluation.  However, it is
-easy to rewrite the assertion and avoid any trouble::
-
-        content = f.read()
-        assert content != '...'
-
 
 assertions about expected exceptions
 ------------------------------------------
@@ -137,6 +120,46 @@ Special comparisons are done for a number of cases:
 
 See the :ref:`reporting demo <tbreportdemo>` for many more examples.
 
+
+Assertion debugging details
+---------------------------
+
+Reporting details about the failing assertion is achieved either by rewriting
+assert statements before they are run or re-evaluating the assert expression and
+recording the intermediate values. Which technique is used depends on the
+location of the assert, py.test's configuration, and Python version being used
+to run py.test.
+
+By default, if the Python version is greater than or equal to 2.6, py.test
+rewrites assert statements in test modules. Rewritten assert statements put
+debugging information into the assertion failure message. Note py.test only
+rewrites test modules directly discovered by its test collection process, so
+asserts in supporting modules will not be rewritten.
+
+If an assert statement has not been rewritten or the Python version is less than
+2.6, py.test falls back on assert reinterpretation. In assert reinterpretation,
+py.test walks the frame of the function containing the assert statement to
+discover sub-expression results of the failing assert statement. You can force
+py.test to always use assertion reinterpretation by passing the
+``--assertmode=old`` option.
+
+Assert reinterpretation has a caveat not present with assert rewriting: If
+evaluating the assert expression has side effects you may get a warning that the
+intermediate values could not be determined safely.  A common example of this
+issue is an assertion which reads from a file::
+
+        assert f.read() != '...'
+
+If this assertion fails then the re-evaluation will probably succeed!
+This is because ``f.read()`` will return an empty string when it is
+called the second time during the re-evaluation.  However, it is
+easy to rewrite the assertion and avoid any trouble::
+
+        content = f.read()
+        assert content != '...'
+
+All assert debugging can be turned off by passing ``--assertmode=off``.
+
 ..
     Defining your own comparison
     ----------------------------------------------
diff --git a/doc/faq.txt b/doc/faq.txt
index 0b76dab26..916d0c001 100644
--- a/doc/faq.txt
+++ b/doc/faq.txt
@@ -47,13 +47,17 @@ customizable testing frameworks for Python.   However,
 ``py.test`` still uses many metaprogramming techniques and
 reading its source is thus likely not something for Python beginners.
 
-A second "magic" issue arguably the assert statement re-intepreation:
-When an ``assert`` statement fails, py.test re-interprets the expression
-to show intermediate values if a test fails.  If your expression
-has side effects (better to avoid them anyway!) the intermediate values
-may not be the same, obfuscating the initial error (this is also
-explained at the command line if it happens).
-``py.test --no-assert`` turns off assert re-interpretation.
+A second "magic" issue arguably the assert statement debugging feature. When
+loading test modules py.test rewrites the source code of assert statements. When
+a rewritten assert statement fails, its error message has more information than
+the original. py.test has a second assert debugging technique. When an
+``assert`` statement that was missed by the rewriter fails, py.test
+re-interprets the expression to show intermediate values if a test fails. This
+second technique suffers from caveat that the rewriting does not: If your
+expression has side effects (better to avoid them anyway!)  the intermediate
+values may not be the same, confusing the reinterpreter and obfuscating the
+initial error (this is also explained at the command line if it happens).
+You can turn off all assertion debugging with ``py.test --assertmode=off``.
 
 .. _`py namespaces`: index.html
 .. _`py/__init__.py`: http://bitbucket.org/hpk42/py-trunk/src/trunk/py/__init__.py