diff --git a/doc/approach.md b/doc/approach.md
index 5fdc705..31879a9 100644
--- a/doc/approach.md
+++ b/doc/approach.md
@@ -1,3 +1,93 @@
-# approach.md
+-[_Courses Open Classrooms_][2]-
 
-_Wait a bit, I am not yet here…_
+# [PyDev] Project 5 : Use public data of [Open Food Facts][1] project
+
+_Last version of this document is avaiable on [github][4]._
+
+## Approach
+
+### Introduction
+
+Search substitute for food product in the [Open Food Facts][1] DB and displays products informations
+
+The user can:
+
+ - interacts with the system in the terminal
+ - saves products for later retrieval
+
+In my opinion, this is a nice cross-techiques project :
+
+- manipulating REST API, with the help of the `requests` package
+- connecting a `MySQL` DB with the help of the `PyMySQL` package
+- working on a (local) `MariaDB` server with the help of the `MySQL` client `mycli`
+- creating an interactive script
+
+The whole project description is avaiable on [OpenClassrooms site][2].
+
+The project is hosted on [github][5].
+
+### Workflow
+
+ - plan the work : features, scripts, tool needed
+ - creates issues for each autonomous package
+ - organize it all in a table
+ - write documentation, as text & (doc)test
+ - writes code
+
+
+### Code construction
+
+To build the script, I followed this approach :
+
+1. take in hand the necessary tools
+2. get started with some the basic features/tools
+3. then add more specificdetailed fonctionalities
+
+Concretely this is expressed by :
+
+- familiarize myself with the data stored in the DB (playing with the CSV file)
+- use the API in a basic way (see `get_product()` function)
+- imagine the main SQL queries
+- using PyMySQL in a basic way
+- create an interface object with the DB
+- create the script interfacing the API and the DB
+- creation of the DB client used by the client
+
+
+### Code organisation
+
+- `config.py`          : configuration
+- `cli.py`             : the user interface script
+- `db.py`              : DB class used by all scripts
+- `function.py`        : files with functions used by all scripts
+- `populate-db.py`     : backoffice script used to populate DB using API
+- `create-db-loff.sql` : backoffice script used to creates local (empty) DB
+- `doc`                : documentation folder
+- `legacy-sql`         : first SQL manipulations (just for memories)
+- `sample`             : OpenFF JSON category & product files (to work offline and lower API usage)
+- `requirements.txt`   : _as itself_
+- `LICENSE`            : _as itself_
+- `README.md`          : _as itself_
+- `.gitignore`         : _as itself_
+
+
+### Difficulties encountered
+
+#### How to interact between the API and the DB
+
+Before exchanging with my mentor I think about using the CSV file as a resource for creating the local DB and then the API will allow to keep it up to date afterwards. This version would have required a large amount of time without bringing a gain of interest.
+
+#### Creating class Db():
+
+Tests are not required in this project but, using [`doctest`][3] is a precious help when I code. Here was an opportunity for me to use them in a class, causing trouble where the test are usefull : In the class declaration or in the methods?
+
+#### Stay in the scope
+
+Not easy to settle for initial specifications that can be interpreted. Similarly for the levels of detail, it is easy to get caught up in a design where they are places for evolutions/new features.
+
+
+[1]: https://fr.openfoodfacts.org/ "Open Food Facts répertorie les produits alimentaires du monde entier"
+[2]: https://openclassrooms.com/fr/projects/utilisez-les-donnees-publiques-de-lopenfoodfacts "Utilisez les données publiques de l'OpenFoodFact"
+[3]: https://docs.python.org/fr/3.6/library/doctest.html "Test interactive Python examples"
+[4]: https://github.com/freezed/ocp5/blob/master/doc/approach.md
+[5]: https://github.com/freezed/ocp5/projects/1