The Telegraph of Claude Chappe -an Optical Telecommunication Network for the Xviiith Century

Claude Chappe (1763-1805) invented a semaphore visual telegraph. The lines between cities were composed by a series of towers (stations), 10-15 km apart, equipped with a pair of telescopes and a semaphore which beams were permitted discrete angular positions. These positions were assigned to numeric symbols in connection with a code book. Where the transmission of a message took days, it only needed tens of minutes with Chappe telegraph (individual symbols may be transmitted at a speed over 500 km/h!). Started during the French Revolution, the network grew to 556 stations covering 3000 miles of lines (5000 km), most of them in France. However, cities like Amsterdam, Brussels, Mainz, Milan, Turin, Venice were also connected. Small networks were also deployed in Algeria and Morocco, while a mobile network was used during the Crimea war. In 1855, it was finally replaced by electric telegraph. The purpose of the presentation is first to present the technology of this optical telegraph, and to demonstrate the modernity of its principles. Nevertheless, political and social implications will be described to show their strong similarities with the (supposed) implications of the World Wide Web. More precisely, such issues as source coding, error detection and signal restoration, control and data signals, routing, regulation and fraud, dissemination of new social behaviours, will be addressed. INTRODUCTION 10 of April 1814, Easter Day, 6 o' clock in the morning, the battle of Toulouse is waged between the French troops (41000 men) of Marshal Soult and the coalition (Portuguese, Spanish and British) troops (52000) of the Marquess of Wellington. There will be no real winner: Soult will be forced to leave Toulouse with his troops the night of the 11, but he will organize an excellent withdrawal like an exercise out of Sandhurst" [1]. Wellington will enter the town the 12, but the British losses are greater than those of the French, and Soult is about to join the army of general Suchet, maintaining a high fighting potential. Therefore useless, about 320 French soldiers and 650 coalition soldiers were killed. Moreover, because communications took a long time, the battle of Toulouse took place four days after the abdication (6 April) of Emperor Napoleon I! Wellington and Soult were of course aware of the imminence of Napoleon's abdication. Messengers (an English colonel and a French colonel) arrived in Toulouse the 12 of April with the great news: the war was over! The messengers had left Paris the 7 of April in the evening. Because of the state of war, they needed nearly five days to reach Toulouse. In normal circumstances, mail from Paris was delivered by postal coach (four days of travel) or mounted couriers (best travel time: three days). In both cases, communications would have been too slow for a cease fire before the battle. However, since August 1794 (20 years ago!), a new communication means had been established in France. It permitted the transmission of information at a speed over 500 km/h, when a postal coach mean speed was around 10 km/h. This new communication means was the optical telegraph of Claude Chappe. Unfortunately, at the fall of the First Empire, the line between Paris and Toulouse was not established yet. The purpose of this paper is to present the technology of this optical telegraph, and to demonstrate the modernity of its principles. Before going into the technical details, it is worth mentioning that the battle of Toulouse could have been the last battle of the First Empire if Napoleon had not decided to return from Elba island. Later will come the episodes of the Hundred Days and of course Waterloo. OPTICAL TELEGRAPHS BEFORE CHAPPE In this paragraph, we will only briefly consider permanent ground systems, operating over a large area, allowing two-way transmission of letters, words or phrases with a single sign. A more comprehensive description of communication systems using messengers, pigeons, mirrors, flags, simple protocol fire beacons etc... used before Chappe telegraph can be found in [2]. The above methods dominated until the advent of Chappe' s invention. The first developments of a telegraph, for which written records exist, is due to Aeneas (350 BC) and Polybius (150 BC). The latter system was able to transmit more sophisticated messages than a simple alarm signal, by encoding the 24 letters of the Greek alphabet into signals using a torch telegraph. Like Chappe's design, transmission of a letter was done in two steps, by transmitting two numbers between 1 and 5. The first number was used to indicate which tablet of a set of five was to be used. Each tablet was labelled with five (or four) letters. The second number was used to indicate which letter was to be read on the selected tablet. Of course, relays forming a chain were used. Surprisingly, it will take nearly twenty centuries before some of these principles would be rediscovered. From Charlemagne to George Washington, the main features of long distance (ground) communications were the same [3]. The invention of the telescope at dawn of the XVII century will foster new designs for telegraphic schemes [2] and will allow day time operation (on the contrary Chappe telegraph will never be operated at night). But the first documented large-scale use will be that of Chappe's, starting a continuous era, with no more gap like the above 2000 year period during which all previous progress was forgotten. This new era started with a period of nearly 60 years corresponding to optical telegraphs dominating long-distance communications. THE INVENTOR Claude Chappe d'Auteroche was born in Brulon, France, in 1763. Claude Chappe initially planned on a career as a member of the clergy, but the French Revolution changed his projects. He then concentrated on scientific work, including long-distance transmission of messages. Most of his work was done with his brothers. They soon rediscovered that complicated messages could be sent using combinations of simple signals. In 1791 a first version of an optical semaphore was devised and successfully used. However, a few more years will be needed to improve the semaphore design and the coding procedure, while efforts were made by the brothers to gain support from the new authorities. In 1794 Claude Chappe was finally put on a government salary. At this stage it is worth remembering that the word telegraph (telegraph, n.means of sending messages; v.send a message) comes from French télégraphe. At first, Claude Chappe wanted to call his invention tachygraphe from the Greek for "fast writer", but he was counselled to decide in favour of télégraphe (from "far writer τηλεγραφω") Claude Chappe committed suicide in 1805, at a time his invention was already a success, to avoid "life's worries" such as criticism and claims from other inventors and competitors. After his death, his brothers Abraham, Ignace and Pierre will be commissioned to organize and chair the telegraph administration. THE TELEGRAPH The mechanical design After early designs using synchronized clocks, and a failed trial to use electricity as a medium for transmission (because no efficient insulator could be found for the electric wires [6]), Chappe devised a semaphore. Preliminary experiments, conducted in 1792, made Claude Chappe and his three brothers convinced that linear arms were more visible in a distance than a shutter semaphore [6] like the one Abraham Edelcrantz was about to built in Sweden in 1794 [2]. Therefore, the final design consisted of a long (4 m x 30 cm) rotating bar (the regulator) with two smaller rotating arms (the indicators) on its ends, counterbalanced with metallic weights. While the regulator could be oriented horizontally, obliquely or vertically, the indicators could be independently oriented in one of seven positions 45 degrees apart, giving a total of 98 combinations. Regulator and indicators were black painted to increase contrast against the sky. Abraham-Louis Breguet, a famous clock maker, designed and built a control mechanism allowing an operator, using a scaled-down model of the semaphore to remotely align the full-scaled one from the inside of a building, using pulleys and ropes. The figure on the right shows an example of semaphore design. It is reproduced from [8]. The identical positions of the semaphore and of the scaled-down model are visible. The two telescopes, each aiming at nearby stations. are not shown. The code It was soon discovered that it was impossible to transmit without using control signals and an efficient coding procedure, because errors were inevitable in the process of transmission. After 1795 and a first use of a signalling code which appeared to need improvements [6], transmissions were done by using 92 combinations of the regulator and indicators. In brief, the regulator could be positioned only vertically or horizontally, and the regulators could be set at angles in increments of 45 degrees, excluding the position where an indicator was extending the regulator. This gave 7 x 7 x 2 = 98 positions, reduced to 92 signals by reserving six signals for special indications. The 92 positions were used to identify in one step a first set (division) of 92 symbols (the alphabet, numbers from zero to nine, some frequently used syllables), or in two steps, first the page number of another 92 page code book (also called a division), and then one symbol (syllables, words, phrases) among the 92 symbols listed on each page. 92 pages time 92 numbered signs, or 8464, means 8464 signs to be transmitted by positioning the semaphore arms twice, transmitting a code pair. After 1799 extra divisions were added giving a total of five. The above six signals reserved for special applications were used to identify the division. The purpose of this system was to save time, by using as few semaphore positions (or symbols) possible to transmit information. It would correspond today to source coding. Basically the formation of a signal was done in two steps and three movements (the French expression "en deux temps et trois mouvements" still meaning today "rapidly done" comes from it). A signal was meaningless as long as the regulator was oblique (left or right), first with the indicators folded in, and then turned to their position (first step consisting of two movements). The left oblique was used for message signals (today the payload) and the right oblique for control signals (today the overhead). The operator had then to verify that the next station was correctly reproducing the signal (corresponding today to restoration at the bit level). This was considered as one of the most important rule to which the operators had to conform. The same error checking was to be made after the second step below. The regulator was then turned to an horizontal or vertical position (second step and third movement). Two examples (from a code table preserved at the Postal Museum in Paris and reproduced in [2]) of the formation of control signals are given in the next table. Each received signal had to be recorded in a book. Additionally, time (hour and minute) was to be recorded for control signals. urgence Paris 1/2 h suspension of activity step 1/first movement the regulators are folders in only if the following station has correctly copied the previous signal third movement step 1/second movement the indicators are set in position by copying the signal of the