Thus, the subject of our interest is a program written in MQL4. Before we start a detailed presentation of the rules of writing programs, it is necessary to describe the basic concepts that characterize a program and its interrelations with information environment. The MetaTrader 4 Client Terminal is known to work online. The situation on financial markets changes continuously, this affects symbol charts in the client terminal. Ticks provide the client terminal with information about price changes on the market.
The Notion of Tick
Tick is an event that is characterized by a new price of the symbol at some instant.
Ticks are delivered to every client terminal by a server installed in a dealing center. As appropriate to the current market situation, ticks may be received more or less frequently, but each of them brings a new quote - the cost of one currency expressed in terms of another currency.
An application operating with the client terminal may work within a long period of time, for example, several days or weeks. Each application is executed according to the rules set for programs of a certain type. For example, an Expert Advisor (EA) does not work continuously all the time. An Expert Advisor is usually launched at the moment when a new tick comes. For this reason, we don't characterize tick as just a new quote, but as an event to be processed by the client terminal.
The duration of Expert Advisor's operation depends on what program code is included in it. Normal EAs complete one information-processing cycle during some tenths or hundredths of a second. Within this time, the EA can have processed some parameters, make a trading decision, provide the trader with some useful information, etc. Having finished this part of its work, the EA goes to waiting mode until a new tick comes. This new tick launches the Expert Advisor again, the program makes its appropriate operations again and returns to the waiting mode. The detailed description of how the appearance of a new tick influences program operation follows below.
The Notion of Control
Speaking about the code execution flow in a program, as well as its interaction with the client terminal, we will use the term of 'control'.
Control is a process of carrying out of actions preset by the program algorithm and the client terminal features. Control can be transferred within the program from one code line to another one, as well as from the program to the client terminal.
Control is transferred in a way similar to that of giving someone the floor to speak at a meeting. Like speakers address a meeting and then give the floor to others, the client terminal and the program transfer control to each other. At that, the client terminal dominates. Its status is higher than that of the program, like the authority of the chairman of a meeting is larger than those of an ordinary speaker.
Before the program is launched, the control is under the supervision of the client terminal. When a new tick is received, the client terminal transfers the control to the program. The program code starts to be executed at this moment.
The client terminal, after it has transferred the control to the program, does not stop its operation. It continues working with maximum performance during the entire period of time it is launched on PC. The program can only start operating at the moment when the client terminal has transferred control to it (like the chairman of a meeting controls the meeting all the time it is going on, whereas the current speaker takes the word for only a limited period of time).
After it has completed its operation, the program returns control to the client terminal and cannot be launched by its own. However, when the control has already been transferred to the program, it returns control to the client terminal by itself. In other words, the client terminal cannot return control from the program by itself. Dynamic actions of the user (for example, forced termination of the program) are an exemption.
When discussing the matters of performance and internal structures of programs, we are mostly interested in the part of control that is transferred within a program. Let's refer to Fig. 2 that shows the general nature of transferring control to, from and within a program. Circles shown in the figure characterize some small, logically completed fragments of a program, whereas the arrows between the circles show how control is transferred from one fragment to another.
Fig. 2. Transferring control in a program
A program that has accepted control from the client terminal (the executing program) starts to make some actions according to its inherent algorithm. The program contains program lines; general order of program execution consists in sequential transfer of control from one line to another in the top-down direction. What and according to what rules can be written in these lines will be considered below in all details.
Here, it is only important to emphasize that every logically completed fragment is executed - for example, some mathematical calculations are made, a message is displayed on the screen, a trade order is formed, etc. Until the current fragment of the program is executed, it retains the control. After it has been fully completed, the control is transferred to another fragment. Thus, control within a program is transferred from one logically completed fragment to another as they are executed. As soon as the last fragment is executed, the program will transfer (return) control to the client terminal.
The Notion of Comment
A program consists in two types of records: those making the program itself and those being explanatory texts to the program code.
Comment is an optional and nonexecutable part of a program.
So, comment is an optional part of a program. It means that a ready program will work according to its code irrespective of whether there are comments in it or not. However, comments facilitate understanding of the program code very much. There are one-line and multi-line comments. A one-line comment is any sequence of characters following double slash (//). The sign of a one-line comment is ended by line feed. A multi-line comment starts with the characters of /* and is ended by */ (see Fig. 3).
Full Guide mql4
Articles
0 commentaires:
Enregistrer un commentaire