commit 0a27d5dd12a8ae6b1661882ded4c9a8a239789ff
Author: Benoit Sigoure <tsuna@lrde.epita.fr>
Date:   Mon Aug 27 11:15:47 2007 +0200

    Explain Git's index.
    
    	* course/svn.html (Things You Should Know): Here.

diff --git a/course/ChangeLog b/course/ChangeLog
index bfc1cc9..433b480 100644
--- a/course/ChangeLog
+++ b/course/ChangeLog
@@ -1,5 +1,10 @@
 2007-08-27  Benoit Sigoure  <tsuna@lrde.epita.fr>
 
+	Explain Git's index.
+	* course/svn.html (Things You Should Know): Here.
+
+2007-08-27  Benoit Sigoure  <tsuna@lrde.epita.fr>
+
 	Make valid XHTML 1.1.
 	* course/svn.html: Various XHTML fixes.
 
diff --git a/course/svn.html b/course/svn.html
index 243d256..9a1c50d 100644
--- a/course/svn.html
+++ b/course/svn.html
@@ -173,6 +173,26 @@ You may find it convenient to watch your repository using
 the <code class="g">gitk</code> repository as you go.
 </li>
 
+<li>
+<b>Index.</b>
+Instead of having <code>.svn</code> directories all over the place, Git uses a
+single <code>.git</code> directory at the root of the working tree.  This
+directory contains an <code>index</code> file that serves a similar purpose as
+the <code>.svn</code> directories: it's a cache that contains an efficient
+representation of <code>HEAD</code> you can quickly diff against (among other
+things).  Unlike SVN, this index aka "Current Directory Cache" is also used as
+a staging area to build the next commit.
+
+Unlike SVN, Git does not track file but content.  Think of Git as a tool
+slightly lower level than SVN.  There are, however, higher level tools built on
+the top of Git to track files (such as
+<a href="http://www.procode.org/stgit/">StGit</a>).  As a consequence, a
+commit is seen as a tree that associates file names to their content (called
+"blobs").  The Git repository tracks the way the content changes and not the
+way a single file changes.  The <code>index</code> is used to incrementally
+build the tree of the next commit.
+</li>
+
 </ul>
 
 <h2 id="commit">Committing</h2>