Programming Review of Visual Basic.NET for the Laboratory Automation Industry

Object Orientation

VB.NET represents a significant shift from an essentially procedural-based language (VB6) to one that demands an understanding of Object Oriented (OO) ² programming concepts. Although it is still possible to create applications for the laboratory that hardly make any use of objects, a thorough understanding of the .NET Framework is essential to fully exploit the language.

VB4 was the first version that allowed use of classes in applications, a major improvement at the time. Many people then proclaimed that VB was an OO language. If the standard definition of an OO language was followed, a key requirement of an OO language is Inheritance; however, this was not available in VB4. Inheritance is now a key feature of VB.NET. All of the other OO features, such as Abstraction, Polymorphism, and Encapsulation are present, as they were in the previous versions of VB6 although they are implemented slightly differently as shown in the examples below.

In Code Example 1—Abstraction and Inheritance, there are two classes ThermoReader and TecanReader. Both are inheriting from the base class PlateReader which has a Read method. The Base class is an abstraction of a Plate Reader, the ability to write a class that represents something.

Code Example 1. Abstraction and Inheritance.

Notice in this example how the base class is constructed. The keyword MustOverride is used for the method Read in the PlateReader base class. When ThermoReader inherits the base class, it must implement its own version of the method Read as indicated by the Overrides keyword preceding it.

As both classes are inheriting from the same base class, a method can be created that can run either type of reader, irrespective of the fact that each object is created from a different class, a concept known as Polymorphism.

Code Example 3. Polymorphism.

The ThermoReader class can be extended to include a property that represents where generated data will be stored when a Read has completed even though this property is not in the base class (see Code Example 4—Encapsulation).

Code Example 4. Encapsulation.

Every object that is instantiated from the ThermoReader class will contain the property DataPath and the value of this property will always be contained within the object. Wherever the object is available, the value of the property Data-Path will also be available to as it is encapsulated within the object.

Types

VB.NET is very flexible when working with types because it can behave as both a static and dynamic typed language. ³ Turn on Option Strict in VB.NET by entering the keyword at the top of each class or as a global setting for the project and it becomes statically typed. If this option is off, the language behaves as a dynamic typed language. This choice allows the programmer to decide what level of type checking he/she wants at compile-time. If Option Strict is turned on, type errors are highlighted at edit time. For most work, Option Strict should be on, to reduce type software bugs. However, static typing can be restrictive and there are times when dynamic typing makes life easier. An example is when working with COM (Component Object Model) on which VB6 is based. If Option Strict is turned on, late binding is not possible.

All of the types that were included in VB6, apart from Variants and Currency, are present in VB.NET although they are different in structure. For example, the Integer type is now 32 bit and not 16 bit, plus there is support for unsigned Integers. ⁴ Essentially, everything is an Object, so it is still possible to work with unknown types if required. This replaces Variants in the previous versions of VB. The problem with using Object for unknown types of data is that it is very prone to error as assumptions are made that the data is of a certain type. VB.NET has also introduced Generics to address this issue. ⁵ Code can be written to handle types before knowing what the type will be. The example below is a very simple class that behaves as a collection.

Code Example 6. Generic Collection.

Instead of referring to an explicit type, the placeholder itemType is used. When creating an instance of this class, itemType is set as a type as shown in the example below, that is, instantiating the class to handle Strings.

In Code Example 6—Generic Collection, the function DirectCast is used. It is strong typed and ensures that the data is or inherits from the expected type. ³ There is also the CType function that will convert data to a different type. This is weaker typed than DirectCast and useful for converting, for example, a String to a numeric type, although the runtime performance is slower.

Operating Systems

Because the language is based upon the Microsoft .NET framework, it is limited to creating applications that run on Windows. Web applications are a slight exception, in that they support cross-platform clients, but deployment does require a Windows Web server. The .NET framework is also available in a 64-bit variation. ⁶ All of the functionality in the 32-bit version of the framework is present; the difference lies behind the scenes in the Common Language Runtime. There is no real noticeable difference in working in both environments. More often than not, programs compiled in a 32-bit OS (Operating System) such as Windows XP can be run on a 64-bit OS. As well as desktop applications, programs can be written to run on a PDA or Mobile Phone by making use of the .NET Compact Framework, a subset of the main Framework.

GUI Development Tools

One of the major factors in take up of any language is how good the IDE (Integrated Design Environment) is. The IDE for VB.NET 2005 is used by all the languages that are part of Microsoft Visual Studio 2005 (Fig. 1), these are Visual C#, Visual C++ and Visual J#. ⁷

OPEN IN VIEWER

Figure 1.

Microsoft Visual Studio 2005 IDE

The IDE has form designers for Windows Forms, traditional Windows applications, and Web Forms using ASP.NET pages. Designing and coding Windows Forms and Web Forms has been made as similar as possible so that both types of applications can share components and much of the complexity of coding Web applications is kept hidden. The support for writing code is a major advance in the IDE. When writing code, suggestions are made and syntax errors and warnings are highlighted. There is a library of reusable code snippets that can be dragged into the code window to speed up development. It is also possible to define and reuse your own code snippets.

Debugging has always been well supported in VB and this has been enhanced. One useful addition is the help dialog that will be displayed with trouble-shooting tips and recommendations on what changes need to be made when an error is encountered in the code. All code can be edited at this point and execution resumed. Also included are Tracepoints which are breakpoints that don't break. Instead, they can report the current program status such as a variable value and let the code continue. This is very useful for programs where having a break in the code would create an unrealistic testing environment such as when waiting for a response from an instrument.

However, it is the new features and tighter integration to data that really make the difference over previous versions. For example, it is now possible to test the methods and properties of a class using the Object Test Bench. An instance of the class can be tested without having to write a test harness for all but the most complicated classes. This is particularly useful for applications that work with instrumentation as you can test each device, assuming a class has been written for each, before testing the entire program. Although this feature would not be high on the list on why you should use Microsoft Visual Studio, it is these additions to the IDE that makes it easier to create a program.

Visual Studio 2005 is available in various editions, Express, Standard, Professional, Tools for Office, and Team System. ⁸ The Express Edition, which is free to download from the Microsoft web site, is a cut-down version targeted at novice developers and is an ideal choice if you wish to complete a basic evaluation of what the product has to offer. As we move up through the versions, the price and the available features increase. The high-end products are the Team System Editions. These are supplied individually as Software Architect, Software Programmer, and Software Tester packages or the Team Edition, which is a package of all three versions. The Team System Editions are primarily for large development organizations and are not really suitable for writing the types of applications found in the laboratory. The most suitable choice is the Standard edition, which is the cheapest version available that has all the functionality required to write a wide range of applications. This version also comes with the Microsoft Developers Network help library, a very useful resource for anyone who takes VB programming seriously.

MultiTasking

There are two ways of running code asynchronously in VB.NET, using Delegates ⁹ or Threads. ¹⁰ Most VB programmers will be familiar with the concept of Events because it is an event driven language. For example, when you click a button a click event is raised. Events can be raised from inside a class so that when a particular event occurs, the client that instantiated the object can be notified. Events are still present in VB.NET but are now based upon Delegates. ¹¹ Using Delegates, it is possible to start a method asynchronously, similar to an event, but it can also have a callback raised upon completion, something which Events cannot do. To run a method through a Delegate, in this example the Read method in the ThermoReader class, a delegate must first be declared.

An instance of the delegate can then be created using an instance of the ThermoReader class. The signature of this delegate must be the same as the object it will be working with. In this example, the Read method of the ThermoReader class has no parameters, so neither does the delegate, DelegateRead.

An AsyncCallback object must also be created that passes

in the method that will be called when objThermo.Read has completed. The only proviso is the method has the parameter of the type IAsyncResult.

It is now possible to invoke the delegate which will execute objThermo.Read along with the Asynchronous callback that has just been created.

Code Example 7. Creating an Asynchronous Callback.

BeginInvoke will not be a blocked call and any code immediately after this call would be executed. For example, the other reader, TecanReader, could be started so both readers are multitasked. When the “Read” method completes, the Asynchronous Callback is executed as shown in Code Example 7—Creating an Asynchronous Callback. Delegates are probably one of the most radical changes to VB.NET and it is certainly not the easiest concept to understand. However, it is a powerful feature that can make applications run at a faster speed.

Code Example 8. Invoke Delegate Using the Asynchronous Callback.

The other way to run code asynchronously is to use Threads. There are various ways to implement threads, the simplest being the BackgroundWorker component. The best method to gain complete control over a thread is to use the Thread class. This can be implemented using the following code, which again uses the ThermoReader class from the previous examples.

Code Example 9. Run Method in New Thread.

This code will run the Read method in a separate thread to the calling code. To then wait until the thread has completed, to synchronize with the calling thread, the following code should be used.

There is much more functionality available that allows complete control over the behavior of each thread. It should be noted that although the example shown is relatively simple, writing multithreaded applications that are thread-safe is difficult because threads tend to need to interact with each other. For example, the Read method in the Reader class may contain code that interacts with a resource, hardware, or software that can only be interacted with by one caller at a time. If two or more instances of the Reader class were then run in separate threads, there could be a scenario where all of the threads are trying to access the resource simultaneously. This often leads to a fatal errors such as memory violations, although the IDE does a good job in handling many of these exceptions.

There are various synchronization techniques available to ensure that code is thread-safe. One way is to use a Lock. In Code Example 10—Sync lock the Read Method, a lock is placed around the code that interacts with Eventlog, which in this example is a resource that writes to a data device. The lock around this code prevents more than one thread accessing this code at any one time. If a thread was running this section of code, another thread which was also running this “Read” method would wait when it gets to this point until the first thread has completed this section and released the lock.

Code Example 10. Sync Lock the Read Method.

Thread synchronization is very important when developing multithreaded applications and requires careful planning to ensure it is thread-safe. The main advantage to using threads is the increased speed in running tasks through the application. In the previous code example, we are making provisions for running Plate Readers at the same time as opposed to one after the other. When running multiple plates in an assay, through many devices, the time savings can be significant, fully justifying the extra time and effort required to develop a multithreaded application.

Hardware Interfacing

VB.NET comes with a series of control components that makes it a fairly simple process to integrate hardware. Most devices found in the laboratory are connected to the PC using one of the following types of connection.

RS232/RS485

RS232 and RS485 can be handled using the SerialPort component that is part of the Framework and available in the IDE. The control is then simply dropped onto the form that will be used to send commands, although it is also possible to create an instance of this control without having to place it (sink it) onto a form. For example, if an instrument is used that has a command set that uses the string “HOME” to move the instrument back to the home position, the following code would execute this.

Code Example 12. SerialPort Component.

Ethernet connections can be made by using the System.Net.Sockets namespace in the .NET Framework. The principles behind using sockets in previous versions of VB are pretty much the same in VB.NET. The main advantage now is the ability to use delegates which gives much more flexibility over the asynchronous control of each socket.

XML

It is no surprise that the main way to handle data in VB.NET is by using Extensible Markup language, known as XML. ¹² The .NET framework is supplied with two base classes, XMLReader and XMLWriter. These are the foundation classes for working with XML, making it very easy to write an application that uses XML documents. For example, to write software to reformat data as it is produced from an instrument such as a 2D Barcode Reader, XML would be by far the best choice as seen in Code Example 13–2D Barcode XML Data.

Code Example 13. 2D Barcode XML Data.

Using the Reader class, we can query the attribute of each element in the XML file with the barcode of the tube we want to return its location. This is a very attractive way to control cherry picking applications. An XML Template file, XSL, could be written for the data so that any other application could read and most importantly, understand what the data represented.

Serialization

In any program that uses objects, data is moved around and sometimes it is necessary to save the data. Previously, the best way was to create a load and save method in the class and then write code to read and write all the properties of the class to the required format. To use XML, the Document Object Model available in the Microsoft XML Parser, MSXML, has to be used. With .NET, Serialization is available and once again it simplifies the whole process for persistent data. To achieve this, pass the object as a constructor parameter into a new instance of the Serialization class. Then define the output stream, XML or Text and call the serialize method. All public properties are then serialized to the defined stream. This functionality is bidirectional so it is possible to deserialize into a new instance of the object. A lot of assumptions are made in the process and the object has to be fairly simple, although it can handle properties that are collections. If data has to be handled differently, implement the Serialization interface in the class, and using the relevant stream, specify how to handle the data.

ADO.NET

In business programming, VB has one of the largest user bases and one of the primary reasons for this is because of the built-in tools for database connectivity. In VB.NET, ADO. NET (Abstract Data Object) can be used to connect to a variety of databases such as SQL Server, Oracle, and MS Access. ¹³ Nearly all data handling in the .NET environment uses XML, and ADO.NET is no different. All data sets are moved around as XML and because XML is now an industry standard, any application on any platform that can read XML, can receive data from an ADO.NET enabled application. ¹⁴ Another key improvement is the ability to use typed data sets allowing data to be accessed through typed programming. A problem with database programming has been writing SQL statements at design time and then only finding out there is a problem when running the application. Using typed data sets, where the field names are actually properties of the data set, this problem is removed because the code is checked at compile time.

Example Application

The following example looks at how a small Windows application can be created to control a Thermo Labsystems Multidrop 96/384. This is a simple device to integrate and one of the most common instruments found in the laboratory.

The first step is to look at the required functionality. Use-case analysis would reveal the following methods:

SetPort

Although SetPort has been identified as a method for setting the RS232 port number, it would make more sense for this to be a property. For this example, ignore the fact that it is possible to dispense to different columns and consider only the entire plate. Now create a new windows application project in Visual Studio. The default form should be renamed to frmMultidrop and this will also be the startup form. A new class file is then added and this is called clsMultidrop. This class is the abstraction of the Multidrop instrument and will contain the methods and properties as described above.

Next, create a class for sending commands using RS232. This class will need a method to send the command and another property to set the port number. In this example, another class file will be created to implement this new class even though it could be written in the existing clsMultidrop class file.

Code Example 15. RS232 Serial Port Class.

The SerialPort object is created using an instance of the class directly from the .NET Framework using the System. IO.Ports namespace, instead of using the SerialPort component. The clsRS232 class shows another feature now available and that is the ability to use constructors with parameters. In previous versions, it was only possible to have a single constructor, New. Constructors can now be overloaded so that when an instance of a class is created, the parameters can be defined in the same line of code. An example of this is highlighted, in Red, in Code Example 15–RS232 Serial Port Class.

Now create an instance of the RS232 class in the Multidrop class so that each method can send the relevant command to the Multidrop.

Code Example 16. Multidrop Class.

This code, encapsulated by clsMultidrop will, when instantiated, control the Multidrop (Fig. 2).

OPEN IN VIEWER

Figure 2.

Multidrop Control GUI

To produce a user interface, use the form that already exists in the project. For this example, simply add one button for each function that is available, using input controls to pass in parameters.