components/org.eclipse.papyrus.designer.components.modellibs.core/html/examples/hello-world.html - papyrus/org.eclipse.papyrus-designer - Git at Google

 <!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN">
 <html>
 <head>
 	<meta http-equiv="CONTENT-TYPE" content="text/html; charset=utf-8">
 	<title>Hello world example with Qompass</title>
 	<link rel="StyleSheet"
 		href="../sitestyle.css"
 		type="text/css">
 </head>
 <body>

 <h1>Hello world example with Qompass</h1>

 In the sequel, we outline a simple hello world example using qompass designer. Please note that the example
 can be found via File->New->Example in the category Qompass examples. There are also two cheat-sheets that explain how
 to create your own example from scratch.
 <p>

 As for all Qompass examples, there are three different parts that needs to be modeled
 <ol>
 <li>The software components
 <li>An abstraction of the hardware platform
 <li>A so-called deployment plan, i.e. a set of component instance definitions along with their configuration and
 	allocation of software components to hardware nodes
 </ol>

 In the sequel, we will go through all of these items with a focus on the the definition of the software components.

 <h2>Definition of software components</h2>

 The example consists of two components, detailed below

 <h3>The hello-world component</h3>

 The hello-world component is shown in the following screen shot.

 <center>
 <img src="screenshots/hw/hello-world.png" alt="The Hello World component">
 </center>
 <br>
 The hello world component inherits from a component called <em>CStart</em>. This is an abstract component defined in the
 Qompass core model library. It owns a port (called start) that provides the pre-defined interface <em>IStart</em>. This
 interface has a single operation <em>run</em>. Components that provide an interface via a port need to implement this
 interface. This include ports that are inherited. In our case, the hello world component needs to implement the <em>IStart</em>
 interface, i.e. contain an operation "run" as well as a behavior for this operation.<br>
 Qompass supports the user by automatically creating derived operations for interfaceS provided via a port, i.e. the operation
 signature is copied (and synchronized on demand) from the interface. The developer needs to provide behavior. In UML, behavior
 can be specified either via activities or by embedding code in a programming language (opaque behavior) into the model.
 Currently, Qompass focuses on the latter,in particular in form of embedded C++ code. A simple C/C++ view allows to enter
 the code, as shown in the following screenshot (there is ongoing work concerning a CDT editor integration)

 <p>
 <center>
 <img src="screenshots/hw/c-cpp-view.png" alt="C/C++ view">
 </center>
 <p>

 The code outputs "Hello World" and an additional message that is an attribute of the hello world component. The attribute <em>message</em>
 is declared as a UML String, but it does not have a particular contents on the level of the component definition. We will
 see in the section explaining deployment plans how this contents can be defined.
 <p>
 The specific start port enables a component to get the initial control. Its <em>run</em> operation is called at system
 start-up by a generated boot-loader.

 <h3>A system component</h3>

 All Qompass applications need to define a top-level component, typically called <em>System</em> that aggregates its components,
 in our case the hello-world component.  It is defined by means of a composite class, contained components are
 modeled via attributes (use composite aggregation for these) typed by the contained components. Note that we define a role (or part)
 within a composition, the hello world component declaration is not part of the system class. This role corresponds to an
 instance of a class in the deployment plan below.<br>
 In our simple hello world example, there is a single component only. Thus, there are no connections between components (so called <em>assembly</em>). take a look at the
 <a href="producer-consumer.html">producer-consumer</a> example to see how the assembly can be done.
 <p>
 <center>
 <img src="screenshots/hw/system.png" alt="The System component">
 </center>


 <h2>Definition of an abstract hardware platform</h2>

 In parallel to the software component, the hardware platform or merely an abstraction thereof can be defined. The main objective
 of this step is to define computational nodes on which the software can be allocated. A node isa stereotyped class that defines
 for the moment only quite basic information about a computational node, such as the provided memory and operating system.

 <p>

 The following screenshot defines a platform for the Hello World component. The <em>HWArchitecture</em> class has a similar
 function for the platform as the <em>System</em> class has for the software: it represents the used notes and their interconnections.
 In this simple example, the hardware architecture contains a single node.<br>
 <center>
 <img src="screenshots/hw/platform.png" alt="A platform for the Hello World component">
 </center>

 <h2>Definition of a deployment plan</h2>

 A deployment plan is a collection of instance specifications, stored within a UML package. It has two objectives. The first is
 to configure an <em>instance</em> of a component, i.e. to provide values for configuration attributes. The second is to allocate
 instance to processing resources such as threads or directly to a node of the platform definition.

 In case of the hello world component, the message that should be written is a configuration attribute. Qompass provides a
 dialog that simplifies the configuration of this attribute. The dialog lists instances and their configuration attributes and
 enables the user to provide a value in a text field. On the model level, a corresponding slot and a value specification are
 produced.<br>

 <center>
 <img src="screenshots/hw/configure-instance.png" alt="configure instance dialog">
 </center>

 The allocation is also simplified by means of a custom dialog. The instances of a system are shown by means of a tree structure
 and the user can either allocate leafs or a parent node. In our case, the whole system should be allocated to node "monoNode", as
 shown in the screenshot below. On the model level, allocation is a MARTE relationship.<br>
 <center>
 <img src="screenshots/hw/allocate.png" alt="allocation dialog">
 </center>

 The last screenshot shows the resulting deployment plan, the package containing two instance specifications and the slots along
 with the configured values. You also see the allocation of the system component. Note, that the visualization of all instances
 in a diagram is often not useful for larger applications, since the graphical visualization does not provide additional
 information compared to the tree structure depicted by the dedicated dialog.<br>

 <center>
 <img src="screenshots/hw/deployment-plan.png" alt="resulting deployment plan">
 </center>

 The last step is the code generation which can be executed via the context menu of a deployment plan. The code generation creates an CDT project
 which can be compiled and executed.

 <h2>Compilation, Execution and Debugging</h2>

 <h3>For Windows users</h3>

 If you are using Windows and it is the first time you are running C++ project using Eclipse CDT,
 you can face some problems during compilation, execution and debugging of the generated code.

 First of all, be sure that you have defined the environment variable for your C++ compiler. To do so, open a command line window  (cmd.exe) and type "gdb".
 You should see an output like below.

 <center>
 <img src="screenshots/cmd_without_error.png" alt="command line without error">
 </center>

 If you do not see this output and see the error below,

 <center>
 <img src="screenshots/cmd_with_error.png" alt="command line with error">
 </center>


 go to "My Computer -> Properties -> Advanced System Parameters". Edit the variable named "PATH" and add the following to the end of its value ";C:\MinGW\bin\"
 <br><br>
 Moreover, another problem you can face can be having no console output after a successful run of the code. In this case,
 <br><br>
 Right-click on your project. Select "Properties".
 <br><br>
 Select the "Run/Debug Settings" Property on the left of the new window.
 <br><br>
 In the right window, click on your executable to highlight (ie - Test.exe) and click "Edit".
 <br><br>
 In the Environment tab, hit "New"
 <br><br>
 Name: <b>PATH</b>
 <br>
 Value: <b>Path to your MinGW bin directory</b> (For example: C:\MinGW\bin)
 <br><br>
 Click "OK" on all windows to close down.
 <br><br>
 Try running again, it should print output to the screen.
 <br><br>
 <a href="http://stackoverflow.com/questions/3443254/eclipse-cdt-using-mingw-does-not-output-in-console">See this link!</a>


 </body>
 </html>
	<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN">
	<html>
	<head>
	<meta http-equiv="CONTENT-TYPE" content="text/html; charset=utf-8">
	<title>Hello world example with Qompass</title>
	<link rel="StyleSheet"
	href="../sitestyle.css"
	type="text/css">
	</head>
	<body>

	<h1>Hello world example with Qompass</h1>

	In the sequel, we outline a simple hello world example using qompass designer. Please note that the example
	can be found via File->New->Example in the category Qompass examples. There are also two cheat-sheets that explain how
	to create your own example from scratch.
	<p>

	As for all Qompass examples, there are three different parts that needs to be modeled
	<ol>
	<li>The software components
	<li>An abstraction of the hardware platform
	<li>A so-called deployment plan, i.e. a set of component instance definitions along with their configuration and
	allocation of software components to hardware nodes
	</ol>

	In the sequel, we will go through all of these items with a focus on the the definition of the software components.

	<h2>Definition of software components</h2>

	The example consists of two components, detailed below

	<h3>The hello-world component</h3>

	The hello-world component is shown in the following screen shot.

	<center>
	<img src="screenshots/hw/hello-world.png" alt="The Hello World component">
	</center>
	<br>
	The hello world component inherits from a component called <em>CStart</em>. This is an abstract component defined in the
	Qompass core model library. It owns a port (called start) that provides the pre-defined interface <em>IStart</em>. This
	interface has a single operation <em>run</em>. Components that provide an interface via a port need to implement this
	interface. This include ports that are inherited. In our case, the hello world component needs to implement the <em>IStart</em>
	interface, i.e. contain an operation "run" as well as a behavior for this operation.<br>
	Qompass supports the user by automatically creating derived operations for interfaceS provided via a port, i.e. the operation
	signature is copied (and synchronized on demand) from the interface. The developer needs to provide behavior. In UML, behavior
	can be specified either via activities or by embedding code in a programming language (opaque behavior) into the model.
	Currently, Qompass focuses on the latter,in particular in form of embedded C++ code. A simple C/C++ view allows to enter
	the code, as shown in the following screenshot (there is ongoing work concerning a CDT editor integration)

	<p>
	<center>
	<img src="screenshots/hw/c-cpp-view.png" alt="C/C++ view">
	</center>
	<p>

	The code outputs "Hello World" and an additional message that is an attribute of the hello world component. The attribute <em>message</em>
	is declared as a UML String, but it does not have a particular contents on the level of the component definition. We will
	see in the section explaining deployment plans how this contents can be defined.
	<p>
	The specific start port enables a component to get the initial control. Its <em>run</em> operation is called at system
	start-up by a generated boot-loader.

	<h3>A system component</h3>

	All Qompass applications need to define a top-level component, typically called <em>System</em> that aggregates its components,
	in our case the hello-world component. It is defined by means of a composite class, contained components are
	modeled via attributes (use composite aggregation for these) typed by the contained components. Note that we define a role (or part)
	within a composition, the hello world component declaration is not part of the system class. This role corresponds to an
	instance of a class in the deployment plan below.<br>
	In our simple hello world example, there is a single component only. Thus, there are no connections between components (so called <em>assembly</em>). take a look at the
	<a href="producer-consumer.html">producer-consumer</a> example to see how the assembly can be done.
	<p>
	<center>
	<img src="screenshots/hw/system.png" alt="The System component">
	</center>


	<h2>Definition of an abstract hardware platform</h2>

	In parallel to the software component, the hardware platform or merely an abstraction thereof can be defined. The main objective
	of this step is to define computational nodes on which the software can be allocated. A node isa stereotyped class that defines
	for the moment only quite basic information about a computational node, such as the provided memory and operating system.

	<p>

	The following screenshot defines a platform for the Hello World component. The <em>HWArchitecture</em> class has a similar
	function for the platform as the <em>System</em> class has for the software: it represents the used notes and their interconnections.
	In this simple example, the hardware architecture contains a single node.<br>
	<center>
	<img src="screenshots/hw/platform.png" alt="A platform for the Hello World component">
	</center>

	<h2>Definition of a deployment plan</h2>

	A deployment plan is a collection of instance specifications, stored within a UML package. It has two objectives. The first is
	to configure an <em>instance</em> of a component, i.e. to provide values for configuration attributes. The second is to allocate
	instance to processing resources such as threads or directly to a node of the platform definition.

	In case of the hello world component, the message that should be written is a configuration attribute. Qompass provides a
	dialog that simplifies the configuration of this attribute. The dialog lists instances and their configuration attributes and
	enables the user to provide a value in a text field. On the model level, a corresponding slot and a value specification are
	produced.<br>

	<center>
	<img src="screenshots/hw/configure-instance.png" alt="configure instance dialog">
	</center>

	The allocation is also simplified by means of a custom dialog. The instances of a system are shown by means of a tree structure
	and the user can either allocate leafs or a parent node. In our case, the whole system should be allocated to node "monoNode", as
	shown in the screenshot below. On the model level, allocation is a MARTE relationship.<br>
	<center>
	<img src="screenshots/hw/allocate.png" alt="allocation dialog">
	</center>

	The last screenshot shows the resulting deployment plan, the package containing two instance specifications and the slots along
	with the configured values. You also see the allocation of the system component. Note, that the visualization of all instances
	in a diagram is often not useful for larger applications, since the graphical visualization does not provide additional
	information compared to the tree structure depicted by the dedicated dialog.<br>

	<center>
	<img src="screenshots/hw/deployment-plan.png" alt="resulting deployment plan">
	</center>

	The last step is the code generation which can be executed via the context menu of a deployment plan. The code generation creates an CDT project
	which can be compiled and executed.

	<h2>Compilation, Execution and Debugging</h2>

	<h3>For Windows users</h3>

	If you are using Windows and it is the first time you are running C++ project using Eclipse CDT,
	you can face some problems during compilation, execution and debugging of the generated code.

	First of all, be sure that you have defined the environment variable for your C++ compiler. To do so, open a command line window (cmd.exe) and type "gdb".
	You should see an output like below.

	<center>
	<img src="screenshots/cmd_without_error.png" alt="command line without error">
	</center>

	If you do not see this output and see the error below,

	<center>
	<img src="screenshots/cmd_with_error.png" alt="command line with error">
	</center>


	go to "My Computer -> Properties -> Advanced System Parameters". Edit the variable named "PATH" and add the following to the end of its value ";C:\MinGW\bin\"
	<br><br>
	Moreover, another problem you can face can be having no console output after a successful run of the code. In this case,
	<br><br>
	Right-click on your project. Select "Properties".
	<br><br>
	Select the "Run/Debug Settings" Property on the left of the new window.
	<br><br>
	In the right window, click on your executable to highlight (ie - Test.exe) and click "Edit".
	<br><br>
	In the Environment tab, hit "New"
	<br><br>
	Name: <b>PATH</b>
	<br>
	Value: <b>Path to your MinGW bin directory</b> (For example: C:\MinGW\bin)
	<br><br>
	Click "OK" on all windows to close down.
	<br><br>
	Try running again, it should print output to the screen.
	<br><br>
	<a href="http://stackoverflow.com/questions/3443254/eclipse-cdt-using-mingw-does-not-output-in-console">See this link!</a>


	</body>
	</html>