diff --git a/CHANGES b/CHANGES
new file mode 100644
index 0000000000000000000000000000000000000000..a9c52160e3dd244a94bd6148b627755250158485
--- /dev/null
+++ b/CHANGES
@@ -0,0 +1,49 @@
+; -*-org-*-
+
+* Version 1.0b
+** The acgc.opt/acgc compiler:
++ Now outputs an "file.acgo" file when compilation is successful
++ Can declare a lexicon as the composition of two other lexicons
+  using the following declaration:
+  lexicon lex_name = lex_name2 << lex_name1
+
+** The acg.opt/acg interpreter:
++ Can load an acg object file using the command:
+
+  load o file.acco;
+
++ The "analyse" command is deprecated. It is replaced by:
+  + A "check" command, prefixed by signatures, that typecheck the
+    typing assignment of a term
+
+    Sig1 Sig2 check term:type;
+
+  + A "realize" command that must be preceded by lexicons and followed
+    by the type assignment of a term. It checks that the term is well
+    typed in the abstract signatures of the lexicons and compute its
+    realizations through the lexicons.
+
+    Lex1 Lex2 realize term:type;
+
++ A "parse" command has been added. It must be preceded with the name
+  of a lexicon and it returns the antecedent by this lexicon of the input
+  term with respect to some distinguished (atomic) type:
+
+  Lex parse object_term:distinguished_type;
+
++ A "query" command has been added. It follows the "parse" command
+  syntax and it outputs the associated query and associated extensional
+  database.
+
+  Lex query object_term:distinguished_type;
+
++ A "idb" command has been added. It must be preceded with the name of
+  a lexicon and it returns the intensional database associated with
+  the lexicon
+
+  Lex idb;
+  
+
+** Current limitation:
++ Only parse images of atomic types
++ The result of parsing with non-linear lexicons is unspecified yet
diff --git a/TODO b/TODO
index 5bb0ba8c80be3267c8dd35c2e4927b43607468d4..66f4ac6b9d3361b6ac48a148f8103fd0413e597c 100644
--- a/TODO
+++ b/TODO
@@ -3,13 +3,13 @@
 * Général
 
 ** Avant Prochaine release
-+ [ ] merge avec trunk
-+ [ ] remplacer tous les gforge.loria en gforge.inria
-+ [ ] réfléchir si changement de construction des numéros de version
-+ [ ] renommer s_datalog et datalog en datalog.prover et datalog.solver
++ [X] merge avec trunk
++ [X] remplacer tous les gforge.loria en gforge.inria
++ [X] réfléchir si changement de construction des numéros de version
++ [X] renommer s_datalog et datalog en datalog.prover et datalog.solver
 + [X] modifier la documentation
 
-** TODO Faire un Bolt package
+** DONE Faire un Bolt package
 
 
 ** DONE Utiliser Bolt (http://bolt.x9c.fr/)
@@ -84,6 +84,8 @@
 + [X] Définir le parsing de la composition de lexique
 + [ ] Définir le parsing pour les types non atomiques
 + [ ] Définir le parsing pour les almost linear terms
++ [ ] Vérifier que les Signature.expand_term ne sont pas redondants
+  avec les Lambda.normalize ~id_to_term
 
 * Divers
 + [X] ajouter la commande compose dans les fichiers de définition de
@@ -92,3 +94,5 @@
 + [X] faire des dumps des interfaces et des représentations binaires
   pour charger plus vite
 + [ ] Ajouter un -nooout option pour acgc ?
++ [ ] Vérifier ce qui se passe avec les différents parenthèsages lors
+  du parsing