Index: libs/serialization/doc/tutorial.html
===================================================================
RCS file: /cvsroot/boost/boost/libs/serialization/doc/tutorial.html,v
retrieving revision 1.3
diff -u -u -r1.3 tutorial.html
--- libs/serialization/doc/tutorial.html	31 Aug 2004 04:06:56 -0000	1.3
+++ libs/serialization/doc/tutorial.html	28 Sep 2004 09:06:48 -0000
@@ -512,6 +512,22 @@
 are met with this system. The <a href="../example/demofile.txt"
 target="test_file">contents of the archive file</a> can
 also be displayed as serialization files are ASCII text.
+<h4><a name="xml-archives">XML Archives</a></h4>
+<p>
+The Boost Serialization library also provides archive classes that generate and read XML
+files - <code style="white-space: normal">xml_oarchive</code> and
+<code style="white-space: normal">xml_woarchive</code> for saving and
+<code style="white-space: normal">xml_iarchive</code> and
+<code style="white-space: normal">xml_wiarchive</code> for loading. They are used exactly
+like any other archive class except that you need to associate the data with the
+corresponding XML tag names.
+</p>
+<p>
+This is accomplished by serializing <a href="wrappers.html#nvp">name-value pairs</a>.
+If used with any of the non XML based archive implementations, the name-value pairs are
+optimized away. It is therefore reasonable to always serialize name-value pairs if you
+imagine your data may ever be serialized into an XML archive.
+</p>
 <hr>
 <p>Revised
   <!--webbot bot="Timestamp" S-Type="EDITED" S-Format="%d %B, %Y" startspan -->