X-Git-Url: http://info.iut-bm.univ-fcomte.fr/pub/gitweb/simgrid.git/blobdiff_plain/0f5e8daaa6e9f74521068aa75837200bcd182ea6..fb1c207eca42103337f6c88ec23603a6ca988fed:/include/xbt/queue.h

diff --git a/include/xbt/queue.h b/include/xbt/queue.h
index 9084a5c4de..3b83e7f99c 100644
--- a/include/xbt/queue.h
+++ b/include/xbt/queue.h
@@ -16,12 +16,20 @@
 SG_BEGIN_DECL()
 
 /** @addtogroup XBT_queue
-  * @brief Synchronized message exchanging queue.
+  * \brief Synchronized message exchanging queue.
   *
-  * These is the classical producer/consumer synchronization scheme, which all concurrent programmer recode one day or another.
+  * These is the classical producer/consumer synchronization scheme, 
+  * which all concurrent programmer recode one day or another.
+  *
+  * The good thing of this implementation is that it works seamlessly
+  * in your universe. When using one of the classical simulation
+  * interface (such as MSG), it achieves the synchronization on top
+  * of the simulator. If you use instead the real life implementation
+  * comming with GRAS, it uses the synchronization of your OS
+  * (whatever could it be). The choice is done at link time.
   *  
   * For performance concerns, the content of queue must be homogeneous, 
-  * just like dynars (see the \ref XBT_dynar section). Indeed, queues use a 
+  * just like dynars (see the \ref XBT_dynar section). Actually, queues use a 
   * dynar to store the data, and add the synchronization on top of it. 
   * 
   * @{