This abstract will describe my three-month training period. I perform this mission in the ORT Group. The main subject will be to improve the search of a company by its activity.

Later, you will see a brief description of the company graced with a report of the division where I will integrate. Following, there will an analysis of my realization and finally you will conclude with a critique of the complete work.


ORT Group is a French society and was created in 1973. It is ranking at the first place for electronic information providing. Today it is working with many organizations (Registre National du Commerce, the secrétariat Général du Gouvernement, the Journaux Officiels and the assemblée Nationale…) to supply an access to the French and common market law to professionals and citizens.

The chairman and managing director of the ORT Group, M. GALANDE, and other peoples are in control of many companies (L'Européenne de données - La Librairie Electronique - Informa - Orchid Telematics Limited - Euro DB). ORT is a limited company with a capital of 35 000 000 F.


ORT owns a great computer park. It is compose of 3 multiprocessors SUN ENTERPRISE 10000 (48 processors for each server, 60 Go of RAM…) and 2 multiprocessors SUN ENTERPRISE 5000. This set enables many types of connection (Internet, Special Link, Minitel) and a lot at the same time (up to 80 000 connection per month).


This infrastructure allow to integrate many standards of the market:

The division where I was integrated is the 'Information retrieval'. Its main activity is WEB application development and Minitel services. It is locating at the company's head in Touraine.

Let's go to see what I have realized during this training period.


This three months were very instructive due to many projects I have to realize. People were trust in my capacity and let me take decision.


During this period I have realized two projects:

Now you will see in detail each part of the training period development.


TThe first part is to offer many solutions to pay on-line by Internet to access ORT information. So I've realized a model enabling: 1. Access without subscription:

2. Access with subscription: To simplify maintenance and developing, ORT uses the service SIPS provide by ATOS: 1. The merchant chosen its mode of regulation: 2. The merchant chosen also conditions of the financial transaction (currency). 3. SIPS is compatible with: The purpose of the model is to: The model was accepted and validated by the supervising team projects. The functioning is explicit. It remains to put it in place on the ORT server. The developer attend to realize the connection with the ATOS Company, therefore it is the beginning of the installation of the service for an on-line usage.

Now you will view the second part of my project.


Before, describe the realization we have to look at the computer organization of ORT. ORT proceeds with an amount of data. It receive many data composed of flat files (word, excel, dos file…) that are convert into ORACLE database (enterprise data) and FULCRUM database (legal and document data).

When these database are used one important criteria is the activity of a company. Activity are saving with a NAF code (Nomenclature d'activités Françaises) normalize by INSEE. NAF corresponding to a definition attach to a code (27.2A = Fabrication de tubes en fonte).


But in databases there is a reference to the code instead of the definition. So the aim of this project is to build a database enable the search of products (like 'voiture') and to obtain a list of NAF code corresponding to this product (502Z = Entretien et réparation de véhicules automobiles, 711Z = Location de véhicules automobiles…).


I have realized this database with DOS files supplied by INSEE but I should respect model of data insertion and parse files with PERL (Practical Extraction and Report Language) and Shell.


I have created a program in C to dialog with the database. To do this, I have used the SearchServer API. Then, a graphical interface was integrated (this interface was programmed with Xforms library).

The interface allow to sail in the NAF code arborescence and following remit:

For this part a second interface was developed for the WEB.

The ORT Group has made the acquisition of OPENWEB. This product is destined for facilitate the installation of complex WEB applications. It is based on the CGI norm and interface on all type of HTTP server using this norm.

OPENWEB is composed of 3 modules:

After being entertained on the OPENWEB functioning and the Javascript programming, I have realized an interface allowing researches on key words on different columns.

For this part have been added, products matching to each NAF. It is therefore in the basis that the largest changes have been undertaken. But to make emerge this contribution of precision it has been necessary to display products in the table and to make under parts in the table to each produce.


In conclusion, it results for the database that simple research but efficient that offer to visualizing the result. The display puts in bold words that have permit selecting the activity.

Therefore for this part the notebook of costs is filled.

See now the part on the graphic interface realized under X-Windows.


The graphic interface under UNIX helps me to become familiar with the FULCRUM SearchServer environment and to discover Xforms library. The goal of this interface was to sail in the activity nomenclature, but I have added a module allowing to research directly words, and even to question others database FULCRUM.

The WEB model is the outcome of my training period, and corresponds to specifications. I would like to see it is implementation on the other site.

For the moment these realizations are part of research within the ORT Group, but I hope that my work will serve for production use and will allow a more convivial access to the research on data.