• the scientific data processing, which consists in helping the planning engineers in the fields of industrial engineering to design and dimension equipment using calculation programmes: nuclear reactors, planes, cars (languages often employed: historically FORTRAN, competed with more and more by C and C++). The scientific data processing is especially used in the research departments and the companies of industrial engineering because it makes it possible to simulate scenarios in a fast and reliable way. The Scuderia Ferrari was equipped in 2006 with one with the most powerful calculators with the world in order to permit the numerical tests of its Formule 1 and to accelerate the development of its prototypes.

• data processing real-time: it consists in defining the driving softwares of systems in direct catch with the physical world: historically initially in aeronautics, the space one, the armament, the nuclear power, but maintaining universally widespread with the miniaturization of the circuits: car, washing machine, etc

• the engineering of knowledge (in English knowledge management ): it is about a data-processing form of engineering which consists in managing the processes of innovation, in all the fields, according to models rather different from those hitherto employed in business computing. Perhaps this form of engineering will make it possible to better put in coherence the three fields management, real-time, and scientist in the organization of the companies. It is interested more in the contents and quality of the databases and of knowledge that in automation of the treatments. It develops already much in the United States, but this is not yet completely perceived in France.

• It is finally necessary to quote the applications of the information ( intelligence in English) economic and strategic, which call upon information technologies, in particular in the analysis of the context, for the search for information (search engines). In addition, with a view to sustainable development, it is necessary to structure the relations with the Recipients, which calls upon other techniques such as the protocols of exchange and the engines of rules.

History

See also: History of data processing

Origins

In 1642, Blaise Pascal also produced a calculating machine mechanical which was for its part marketed and whose nine specimens exist in museums like that of the Arts and trades and in private collections (IBM).

The late discovery of the Machine of Anticythère watch that the Greek of Antiquity themselves had started to produce mechanisms of calculation in spite of their reputation of general contempt for the technique (contradicted besides by work of Archimedes).

However, it will be necessary to await the definition of the concept of Programmation (illustrated in first by Joseph Marie Jacquard with its weaving looms with perforated cards, followed Boole and Ada Lovelace as regards a theory of the Programmation of the mathematical operations) to have a base making it possible to connect elementary operations in an automatic way.

Data processing

Science of the numbers and Numbering system

• Of the binary system invented per G. Boole, used by the computers with the restricted enumeration of the 1-2-3 to say much from 4 of use at the primitive ones, the men used numerical systems which indicate the degree of culture or vigilance of the people auprès of which they were in application.

• Sumériens used the sexagesimal system still used nowadays to measure the hour which counts 60 minutes divided into 60 seconds. It was taken again by the Greeks for their astronomical calculations, in the calculation of the angles and time.

• the Romans made use of their 10 fingers and practiced the decimal system to constitute centuries legionaries. They marked the thousands by a vertically barred circle. Deformed this sign gave the “M” to indicate 1.000 and half of this symbol for “D” to indicate 500. The decimal system in application at the Romain S without the knowledge of the Arab numerals, did not facilitate to the task Arithmétique intendants charged to make the accounts.

• the Celts for their part went until using in more the ten toes, which widened their numerical system with 20. The last Celts on the continent, nowadays, still use this system to appreciate all the quantitative values of the everyday life.

• At the time of a safari hairs, feathers and with the phacochères in Senegal, one can discover the system quincal used by Sérères. Why emanating from Sérères? Because Sérères incarnate the tribe of hunters to the Senegal. The population sénégalaise includes/understands four principal ethnos groups. In addition to Sérères, one meets of Wolofs the dominant ethnos group there, their language being recognized national language and French acts as official language. Toucouleurs and Peuls form the two other ethnos groups. On Sérères fell traditionally the role of providers of Gibier with regard to the others Ethnie S, they of agricultural and pastoral culture. By hordes, Sérères carried out displacements by driving out village in village armed with short sticks, arcs and lances. The shot down game; guinea fowls, young partridges and pigeons, were it by this stick which they launched as a virtuoso as soon as the tracked game took its take-off in savanna. Phacochères and other animals with shoes were driven out with the arc, with the lance and completed with the sagaie or the spear by the carriers and the Traqueur S. the original numerical system of Sérères is quincal. It is easy to guess it why of this practice. In the execution of their activity, one with the hands only was available for additional occupations. The other remained to always tighten a weapon, the stick, the sagaie, the lance… to count. Den - niet - deny - njar - gurun; has five is made the carryforward: gurun-den etc produces It hunting was exchanged against agricultural produce, textile, all in all against products of first need for these travelling hunters which cultivated the concern of being made respect. The jewels and ceramics occupied a choice place in barter, these ladies occupy still and always at the time of meetings of any kind the first rank not to appear among the misfits.

• the Jews practiced the numeral system most elaborate. By exploiting the 12 phalange S of the eight Finger S deducted using the inches. They instituted the Grosse, a measuring unit still in force for some Marchandise S. One arrives there while making use for example of the inch of the left hand to deduct the phalanges. Arrived at the last phalange from the little finger, 4 × 3, makes it possible to record the first dozen using the inch of the right hand which points on the first phalange of the index of the right hand. This system thus makes it possible to count up to 144.

Clay shelves were used as document with the scribes, sitted on the quays of unloadings for the assumptions of responsibility of the arrivals or in the warehouses at the time of the exit or the entry of the goods. The Romain marked of a feature with 10, the Celt with 20 and the Jew with 144 made at each end of his numeral system to record to the last part the movements to be recorded.

George Boole, mathematician English (1815-1864), was the inventor of the binary system. Without its system, there would be no transistorized computers which function thanks to 0 and of the 1, which makes it possible to go in calculations ad infinitum.

Modern data processing

Data processing is thus a field coldly developed, even if it finds its origins in antiquity (with the Cryptographie) or in the calculating machine Blaise Pascal, at the 17th century. It is only at the end of the Second world war that she was recognized like a discipline with whole share and developed methods, then a methodology which were clean for him.

Its image during some time was overrated : because the first to program computers had been engineers broken with the technique of the differential equations (the first computers, scientists, were used much for this purpose), of the programmers without particular formation, sometimes besides resulting from data processing, they readily sought to profit also from this label of rocket scientist in order to justify wages made comfortable by:

• the high price of the computers of the time (amounting in what would be tens of million euros today taking into account the Inflation, it relegated to the second plan the considerations of parsimony on the wages);
• the aspect presented like not very accessible from their discipline and a mathematical myth of difficulty maintained around. In fact, the first computers did not program in a way very different from that of the programmable computers used today in the colleges and colleges, and controlled by some fourteen year old pupils but the field was new and the algorithmic one requires a certain degree of concentration associated, perhaps wrongly, with the pure reflection.

The emergence of a really scientific aspect in the programming itself (and not in the only scientific applications that one programs) off appears only with the series The Art Computer Programming of Donald Knuth, professor with the Université of Stanford, the end of the Années 1960, still unfinished monumental work in 2004. Work of Edsger Dijkstra, Niklaus Wirth and Christopher Strachey proceeds of approach also very systematic and it such a quantified.

One asked Donald Knuth in the years 1980 if it were according to him to better attach data processing ( computer science ) to the Electronic engineering - what is often the case in the American universities - or to a department of Mathématiques. He answered: “ I would readily classify it between the Plomberie and the breakdown service Automobile ” to still underline the side Artisan Al of this young person Science.

However, the strong scientificity of the first three volumes of its encyclopedia suggests that it is rather a joke of its share. Moreover, control of languages like Haskell, Ocaml or even APL request a level of abstraction all the same nearer to that of mathematics than of the two quoted disciplines.

The miniaturization of the components and the reduction of the production costs, associated with a need increasingly pressing for data processing of all kinds (scientific, financial, commercial, etc) involved a diffusion of data processing in all the layers of the economy like life of the every day.

In France, data processing started with really developing only in the years 1960, with the Calculation Plan. Many things were known as from this point of view. As often in History, there can be mistakes in interpretation.

Approaches functional

Like statement above, data processing is the automated treatment of data by an electronic device: the computer; the German-speaking ones speak about elektronische Datenverarbeitung/EDV (“electronic treatment of data”), the english-speaking of information technology/IT (“information technologies”), i.e.:

• given or information : in fine , the computer handles continuations or packages of 0 and 1 which are used to represent all kinds of objects:
• of… the numbers obviously, in the case of scientific computations (floating) or accountants (decimal, or binary entirety)… ;
• a text, letters (character S), which one can format with a Word processing, to print, to send by Email…;
• of the vectorial Drawing (CAD, software of illustration, and typography);
• of the Image S statics (Photography S) or animated (Video), of the holograms;
• of the its S, recorded (technical of the direct to disk ) or manufactured by the computer (Synthétiseur), that they are sound effects, music (cf Musique and data-processing) or word;

• automated : the user does not intervene, or little, in the data processing; the treatment is defined in a program which is held all alone, the user is satisfied to provide parameters of treatment; the automatic routine proceeds according to a algorithm, the establishment of this program is the field of the programming.
• treatment: these data are:
• created:
• numbers: automatic acquisition of data of an experiment with a computer;
• text: to type a text with the keyboard;
• images: drawings carried out with the mouse or on a graphics tablet, synthesis of image (to present a project -   fictitious object in the course of conception  -, medical imagery, artistic drawing -   Computer graphics   -, Cartoon film or Pixilation) or digitalization of an existing image (scanner, numerical camera) or animated images (Numeric camera, Webcam);
• sounds recorded (Microphonous) or recreated starting from a virtual partition (Synthetizer) or of a text (Voice synthesis).
• analyzed:
• numbers: the analysis of the numbers comes under the field concerned (mathematics, physics, economy…) ;
• text: to seek the occurrences of words in a text to draw some from the statistics, helps with the orthographical and/or grammatical correction, and, more generally, automatic treatment of the languages (TAL);
• images: one can want to identify an object (pattern recognition, recognition of the characters or OCR), or to determine the surface covered by a color (for example to quantify a covered surface);
• sounds: analyzes spectral, voice recognition.
• modified:
• numbers: calculations;
• text: modification of an existing text, automatic translation in another language (or computer programming language);
• images: modification of contrast, the luminosity, the colors, effects special;
• sounds: application of effects (reverberation, distortion, adjustment height);
  comme there exists, according to the programs and the needs, a large variety of possible codings to represent each type of information, much of treatments consist in converting the data of a format towards another…
• filed then restored:
• the means and techniques of filing vary according to and the desired quantity shelf life of data concerned: electronic memories, magnetic bands, disks magnetic or optical;
• the means of restitution depend on the nature of the data: screens or printers for the text and the images, loudspeakers or instruments MIDDAY for the sounds…

Organisational approach

Data processing for the organization is an element of a system of treatment of information (the entries can be forms paper for example) and of automation. Since Henry Ford, the automation of the tasks having been identified like a Competitive advantage, the question is: what can one automate?

As much it is relatively easy to automate manual duties, as much it is difficult to automate the professional work and sometimes creative. The approach of data processing in an organization thus starts with the elucidation of the processes, i.e. the modeling of the trade. After validation, the MOA (Control of Work) provides the functional specifications of (the work) which will be used as reference in the design for the MOE (Maîtrise of work).

This design will then be carried out in the respect of a Development cycle which defines the roles and responsibilities for each actor. Thus, the exchanges between MOA and MOE are not summarized with the control of the building sites (held of the deadlines and the costs, and validation of deliverable), the MOA and the MOE are guarantors (possibly responsible on a legal level) of the coherence of the information systems, and adequacy of the IT solution with the problems initially noted end-users.

Material

See also: Computer material

One also uses the English term hardware (literally “hardware”) to indicate the Computer material. They are all the components which one can find in:

1. The Computer S and their peripheral: a Ordinateur is a whole of electronic circuits making it possible to handle data in binary form , represented by variations of electric signal . There exist various types of Ordinateur S:

• the microcomputers.

• the work stations.

• the Mainframe S.

• the Waiter S.

• PDA ( Personal DIGITAL Assisting , still called organiseurs ).

• the Center of Media ( Media Center ).

• And many other apparatuses.

Software

See also: Software

The software indicates the part at first sight immaterial of data processing, the organization and the data processing: programs. One indeed quickly realized that very advanced machines technically for their time, like the Bull Gamma 60, remained unsaleable as long as there were not programs to deliver to make them immediately operational. IBM launched between 1968 and 1973 a kind of ancestor of the Free software with its computer 1130, policy which ensured this one by snowball effect an immediate and planetary success, but the conclusions of a lawsuit Antitrust prohibited to him to distribute software voluntarily.

The world of the Mainframe S classifies the software of following categories:

• operating systems;
• databases, like DB2, Ingres or Oracle;
• programs of communication, like CPC or RSCS;
• teleprocessing monitors;
• compromise systems, like CICS or OpenUTM;
• systems of Time-sharings, used for calculation or the development;
• Compiler S translating the languages into instructions machine and calls system;
• all the remainder entered in a named category Logiciels applicatifs .

More simply one generally distinguishes three types of software (by order of proximity of the material):

• the Microsoftware
• the Operating system
• software and user applications (in English software )

One classifies also the software in free and owner, although both are sometimes mixed to differing degree. Some have a function Bureautique or Multimédia such as for example the video games. Certain software acquired known names of all.

The core of the Operating system creates the bond between the hardware and the software. A software, when it is provided in its binary form , would be usable only with a given operating system (because it uses the services of them), and would function only on one specific material (because it uses of it the Instruction code). A more recent design, since the medium of Years 1980, consists in distributing the confused binary software all, and providing them with a system of licenses by tokens or tokens allowing the use of NR simultaneous copies of the software on the network, all confused materials. This approach is majority in the world UNIX.

On the initiative of Richard Stallman and GNU, starting from 1985, a mobility of programmers refuses this logic owner and those are moulted as inventive originators to launch out in the development of tools and libraries system free and compatible with the system UNIX. It is however the independent project Linux, initiated by Linus Torvalds, based on work and the tools of the GNU, which will lead in the creation of an operating system complete and free called GNU/Linux.

A good part of the current software function in a graphic Environnement to interact with the user. The diversity of the computing systems revealed a technique aiming at combining best each one of these universes: the emulator. It is about a software making it possible to simulate the behavior of another system in that which one uses,

• either so that a machine seems to be another (see IBM 1130),
• or to simulate the behavior of an operating system (for example DOS or Windows under GNU/Linux).

The English term is software , at the origin a pun between hardware (“hardware”, to indicate the material) and the opposition software / hardware (soft/hard), opposition between the material (the hard one) and the immaterial one (slackness). The French translations material and software return perfectly this opposition and this complementarity.

The software normally fulfills an awaited function of its users. Nevertheless, of the side effects (sometimes named by misinterpretation of translation effects edge ) exist. Sometimes even, certain software is intended to harm, like the Computer virus S, is named in English, by analogy with software : malware (that one could translate by the neologism nuisiciel , or Malicious software ).

The creation of the software

A computing project fits in a Development cycle which defines the great stages of the realization (planning), in the way in which one passes from one stage to the other (incremental model, out of V, spiral, method up, Extreme programming, etc). For the small projects (or the small development teams), this reflection is often neglected (one is distributed the modules and each one develops in its corner). This is a frequent cause of errors (bug S) and of nonconformity (the finished product is not in conformity with the user expectations). But even the enormous projects, with much of means, are victims of this negligence; thus, the failure of the first flight of ARIANE 5 was due to a problem of software, etc a project can then integrate an approach of the quality and Reliability of the computing systems in order to control as much as possible the finished product.

A project includes/understands the following stages (according to the incremental model):

• the establishment of a Schedule of conditions which defines the specifications to which will have to answer the software;
• the definition of the environment of exécution  (data-processing Architecture):
• standard (S) of computer on which the software must function (station of calculation, desktop machine, laptop, personal assistant, cellphone, automated teller of bank, computer embarked in a vehicle;
• standard and version of (of) the operating system (S) subjacent;
• peripheral necessary for the recording of the data and the restitution of the results (graphic storage capacity, random access memory, possibilities…) ;
• natural of connections network between the components (level of confidentiality and reliability, performances, Communication protocols…) ;

• the design of the application and its components, and in particular of the interactivity enters the developed modules: structure of the shared data, treatment of the errors generated by another module…: it is the field of the software Génie;

• the installation of a development strategy:
• allocation of the functions between the developers or the development teams, which will ensure the coding and the tests;
• the test plan of the software, to make sure that it fulfills well the mission for which it was written, under all the conditions of use which he will be able normally to meet, but also in borderline cases.

After each one of these phases, one can have a stage of Recette, where the customer will validate the choices and the proposals of the project superintendent.

The phase of programming consists in describing the behavior of the software using a Computer programming language. A Compilateur is then used to transform this code written in a comprehensible data-processing language by human into a comprehensible code by the machine, the result is achievable. One also can, for certain computer programming languages, to use a Interpréteur which carries out a code progressively its reading, without necessarily creating the achievable one. Lastly, an intermediary consists in compiling the code written towards bytecode . It is also about a binary, comprehensible format only by one machine, but it is intended to be carried out on a virtual machine , a program which emulates the main components of a real machine. Main advantage compared to computer code is portability theoretically increased (it “is enough” to establish the virtual machine for an architecture given so that all the programs in bytecode can be carried out there), portability which made, after its slowness, the reputation of Java. It should be noted that these three procedures are by no means incompatible. For example, OCaml has at the same time of an interpreter, a compiler towards bytecode, and a compiler towards native code for a large variety of processors. Once written (and compiled if necessary), the code becomes a Logiciel.

For projects of great amplitude, requiring the collaboration of many programmers, even of several teams, one often has recourse to a common methodology (for example MERISE) for the design and to a Computer aided software engineering (CASE) for the realization.

During the programming and before the delivery of the finished product, the program is tested in order to check that it functions well (including in cases of use in degraded mode) and that it is in conformity with waitings of the end user. The intermediate tests make it possible to make sure that each module of code fulfills a function correctly: they are the unit tests . The final tests which check the good sequence of the modules and the treatments are integration tests .

For certain applications asking for an high level of reliability, the tests are preceded by a stage of checking , where specialized software carries out (generally on the source code, but sometimes also on the compiled code) a certain number of analyzes to check partially the good performance of the program. It is not however possible (and of the mathematical theorems show why), to guarantee the perfect correction of all software by this means and the phase of test thus remains necessary. It is also supplemented, when it is about an evolution of an existing application, many automated tests of Not-regression. The tests either not being able to guarantee the absence of errors completely, it is good to supplement them by phases of checking per second reading: techniques exist to try to make this checking exhaustive.

Statistical : the creation of a software is a difficult task; approximately 31% of the computing projects are abandoned before being finished, more than 50% of the projects cost the double of the cost initially estimated and only 15% of the projects finish in times and according to the budget défini. The needs for only maintenance for existing can take up to 50% of manpower of a team charged of a software (but, it is there a function painful, ungrateful, little developing and which rejects and démotive often good programmers).

However with the use of methods like method UP or the Extreme programming, these statistics tend to improve. In particular, thanks to a development by successive iterations where the phases of analysis, of design, realization and test are repeated several times during the lifespan of the project and produce an achievable product each time. The customer can after each iteration " tester" the product and to deliver its opinion. These methods thus allow better managements of the costs and especially of quality while strongly reducing the risks of nonconformity with the wishes of the customer.

Data processing

The Information, to be treated, must be:

• represented by a coding:
• one uses a binary numbering system , where the informational unit element is the bit (contraction of English B inary dig' it : binary digit). The bits are generally gathered by eight, to constitute Octet S (or Byte S). A byte can be represented by the sequence of the bits which constitute it (for example: 00101110 ) or by a pair of values Hexadécimal be (for the same example: 2nd ), more compact. The choice of binary does not result from the mystic, but quite simply to use simple circuits of commutation' , which have very broad tolerances and consequently of low costs;
• one represents the structuring of information to allow exchanges between software components and material components. For that, one defines languages and formalisms of representation.
• stored in permanent systems (memories known as of mass ) or not (memories known as volatile ).

Exchanges of information: protocols and standards

A protocol can thus define:

• a language of description of instructions and data graphic (example: AGP);
• a standard of orders and flow of information for a mass memory (examples: SCSI, FireWire, IDE, Serial ATA);
• of the exchanges between the processor and the expander cards (examples: NCV, PCI Express, ISA);
• of the methods of transfer of information between peripherals (example: USB) or on a network TCP/IP, Internet, ATM, X.25);
• of the orders between a customer and a waiter (examples: POP3 , IMAP , HTTP , FTP …) ;
• of the specific computerized exchanges of information (examples: EDI, EAI, X.400, X.500).

Certain protocols are defined by standards to allow the interworking of the materials or software implementing them. Other standards define, always in the field of the exchanges of information:

• of the languages of representation of information without to define the way in which this information can be exchanged (examples: ASN.1 , XML );
• of the network architectures (examples: OSI model, WiFi, Ethernet, Token Ring).

Data storage

As regards storage of information, one distinguishes the device allowing to physically record it (peripheral and component) in the way in which one structure and represents information to facilitate his treatment.

; Mass memory

• File of perforated cards
• magnetic Magnetic band
• Removable disk (Diskette)
• magneto-optical Disc
• Hard drive (magnetic disk embarking the mechanism, electronics and the play-back heads)
• removable Optical disk (CD-ROM, CD-R, CD-RW but also DVD-ROM, DVD-R, DVD-RW, DVD+R, DVD+R DLL, DVD+RW, DVD-RAM, GD-ROM, HD-DVD, Blu-ray)
• nonvolatile electronic Memory (Memory flash used in particular in “ Key USB”)

; Volatile memory

• RAM, (Random Memory Access)

; Organization of the data for storage

• Formats (extensions) of files
• Filesystem
• Database
• Directory

Distribution of materials and software data-processing

Historically data processing was distributed (resold) by the large manufacturers who treated on line with their customers; majority of those being large accounts or public agencies. Progressively of the fall in the prices of the systems, the market widened, obliging the manufacturers to structure themselves for better diffusing their product and to be pressed on partners. These partners were at the beginning mono-mark and often worked in the shape of semi-exclusive agent then they transformed themselves with the wire of time into independent retailers multi-marks.

Today the distribution of the data-processing products is made in the shape of multiple channels of distribution, among which one counts the direct sales, the E-trade, the chains of retailers, the groupings of retailers, the mail-order selling.

The data-processing wholesalers have a key role in the data-processing distribution and are a point of passage quasi obliged for the companies which chose the indirect sale (by a network of retailers). The wholesalers, who they are general practitioners or specialized, address the multitude of small sales outlets or the service companies for which the activity of trade represents a low volume of activity.

France counts approximately 40.000 data-processing companies and télécoms of which nearly 15.000 have an activity of distribution.

It should be noted that if the distributing number of companies of the products data processing do not cease extending that is much more with the profit of the structured distribution aiming at the general public that independent retailers of which the number does not cease decreasing and whose activity turns more and more to the concept of data-processing service company like software firm.

See also the article: Terminology of the data-processing distribution