The Microsoft .NET Framework consists of two elements: a runtime environment called the Common Language Runtime (CLR), and a class library called the Framework Class Library (FCL). The FCL is built on top of the CLR and provides services needed by modern applications.
While applications targeting the .NET Framework interact directly with the FCL, the CLR serves as the underlying engine. In order to understand the .NET Framework, one first must understand the role of the CLR.
The CLR is a modern runtime environment that manages the execution of user code, providing services such as JIT compilation, memory management, exception management, debugging and profiling support, and integrated security and permission management.
Essentially, the CLR represents the foundation of Microsoft's computing platform for the next decade. However, it has been a long time in the making. Its origins can be traced back to early 1997, when products such as Microsoft Transaction Server (MTS) were starting to deliver on the promise of a more declarative, service-oriented programming model. This new model allowed developers to declaratively annotate their components at development time, and then rely on the services of a runtime (such as MTS) to hijack component activation and intercept method calls, transparently layering in additional services such as transactions, security, just-in-time (JIT) activation, and more. This need to augment COM type information pushed the limits of what was possible and useful with IDL and type libraries. The COM+ team set out to find a generalized solution to this problem.
The first public discussion of a candidate solution occurred at the 1997 PDC in San Diego, when Mary Kirtland and other members of the COM+ team discussed a future version of COM centered on something called the COM+ Runtime, and providing many of the services such as extensible type information, cross-language integration, implementation inheritance, and automatic memory management that ultimately resurfaced in the CLR.
 Two sites, http://www.microsoft.com/msj/1197/complus.aspx and http://www.microsoft.com/msj/1297/complus2/complus2.aspx, contain articles by Mary Kirtland on the COM+ Runtime.
Soon after the 1997 PDC, Microsoft stopped talking publicly about the technology, and the product known as COM+ that was released with Windows 2000 bore little resemblance to the COM+ Runtime originally described. Behind the scenes, however, work was continuing and the scope of the project was expanding significantly as it took on a much larger role within Microsoft.
Initially codenamed Lightning, the project underwent many internal (and some external) renamings, and was known at various times as COM3, COM+ 2.0, the COM+ Runtime, the NGWS Runtime, the Universal Runtime (URT), and the Common Language Runtime. This effort ultimately surfaced as the .NET Framework, announced at the July 2000 PDC in Orlando, Florida. During the following 18 months, the .NET Framework underwent an extensive public beta, culminating in the release of Version 1.0 of the Microsoft .NET Framework on January 15, 2002.
To better understand the CLR, consider how compilers that target the .NET Framework differ from traditional compilers.
Traditional compilers target a specific processor, consuming source files in a specific language, and producing binary files containing streams of instructions in the native language of the target processor. These binary files may then be executed directly on the target processor.
.NET compilers function a little differently, as they do not target a specific native processor. Instead, they consume source files and produce binary files containing an intermediate representation of the source constructs, expressed as a combination of metadata and Common Intermediate Language (CIL). In order for these binaries to be executed, the CLR must be present on the target machine.
When these binaries are executed, they cause the CLR to load. The CLR then takes over and manages execution, providing a range of services such as JIT compilation (converting the CIL as needed into the correct stream of instructions for the underlying processor), memory management (in the form of a garbage collector), exception management, debugger and profiler integration, and security services (stack walking and permission checks).
This compilation and execution model explains why C# is referred to as a managed language, why code running in the CLR is referred to as managed code, and why the CLR is said to provide managed execution.
Although this dependency on a runtime environment might initially appear to be a drawback, substantial benefits arise from this architecture. Since the metadata and CIL representations are processor architecture-neutral, binaries may be used on any machine in which the Common Language Runtime is present, regardless of underlying processor architecture. Additionally, since processor-specific code generation is deferred until runtime, the CLR has the opportunity to perform processor-specific optimizations based on the target architecture the code is running on. As processor technology advances, all applications need to take advantage of these advances is an updated version of the CLR.
Unlike traditional binary representations, which are primarily streams of native processor instructions, the combination of metadata and CIL retains almost all of the original source language constructs. In addition, this representation is source language-neutral, which allows developers to build applications using multiple source languages. They can select the best language for a particular task, rather than being forced to standardize on a particular source language for each application or needing to rely on component technologies, such as COM or CORBA, to mask the differences between the source languages used to build the separate components of an application.
Ultimately, the CLR exists to safely execute managed code, regardless of source language. In order to provide for cross-language integration, to ensure type safety, and to provide managed execution services such as JIT compilation, garbage collection, exception management, etc., the CLR needs intimate knowledge of the managed code that it is executing.
To meet this requirement, the CLR defines a shared type system called the Common Type System (CTS). The CTS defines the rules by which all types are declared, defined and managed, regardless of source language. The CTS is designed to be rich and flexible enough to support a wide variety of source languages, and is the basis for cross-language integration, type safety, and managed execution services.
Compilers for managed languages that wish to be first-class citizens in the world of the CLR are responsible for mapping source language constructs onto the CTS analogs. In cases in which there is no direct analog, the language designers may decide to either adapt the source language to better match the CTS (ensuring more seamless cross-language integration), or to provide additional plumbing that preserves the original semantics of the source language (possibly at the expense of cross-language integration capabilities).
Since all types are ultimately represented as CTS types, it now becomes possible to combine types authored in different languages in new and interesting ways. For example, since managed languages ultimately declare CTS types, and the CTS supports inheritance, it follows that the CLR supports cross-language inheritance.
Not all languages support the exact same set of constructs, and this can be a barrier to cross-language integration. Consider this example: Language A allows unsigned types (which are supported by the CTS), while Language B does not. How should code written in Language B call a method written in Language A, which takes an unsigned integer as a parameter?
The solution is the Common Language Specification (CLS). The CLS defines the reasonable subset of the CTS that should be sufficient to support cross-language integration, and specifically excludes problem areas such as unsigned integers, operator overloading, and more.
Each managed language decides how much of the CTS to support. Languages that can consume any CLS-compliant type are known as CLS Consumers. Languages which can extend any existing CLS-compliant type are known as CLS Extenders. Naturally, managed languages are free to support CTS features over and above the CLS, and most do. As an example, the C# language is both a CLS Consumer and a CLS Extender, and supports all of the important CTS features.
The combination of the rich and flexible CTS and the widely supported CLS has led to many languages being adapted to target the .NET platform. At the time of this writing, Microsoft was offering compilers for six managed languages (C#, VB.NET, JScript, Managed Extensions for C++, Microsoft IL, and J#), and a host of other commercial vendors and academics were offering managed versions of languages, such as COBOL, Eiffel, Haskell, Mercury, Mondrian, Oberon, Forth, Scheme, Smalltalk, APL, several flavors of Pascal, and more.
Given the level of interest from industry and academia, one might say that .NET has spawned something of a renaissance in programming-language innovation.
Developer needs (and Windows capabilities) have evolved much since Windows 1.0 was introduced in November 1985. As Windows has grown to meet new customer needs, the accompanying APIs have grown by orders of magnitude over time, becoming ever more complex, increasingly inconsistent, and almost impossible to comprehend in their totality.
Additionally, while modern paradigms such as object orientation, component software, and Internet standards had emerged and, in many cases, joined the mainstream, these advances had not yet been incorporated into the Windows programming model in a comprehensive and consistent manner.
Given the issues, and the degree of change already inherent in the move to a managed execution environment, the time was ripe for a clean start. As a result, the .NET Framework replaces most (though not all) of the traditional Windows API sets with a well-factored, object-oriented class library called the Framework Class Library (FCL).
The FCL provides a diverse array of higher-level software services, addressing the needs of modern applications. Conceptually, these can be grouped into several categories such as:
Support for core functionality, such as interacting with basic data types and collections; console, network and file I/O; and interacting with other runtime-related facilities
Support for interacting with databases, consuming and producing XML, and manipulating tabular and tree-structured data
Support for building web-based (thin client) applications with a rich server-side event model
Support for building desktop-based (thick client) applications with broad support for the Windows GUI
Support for building SOAP-based XML web services
The FCL is vast, including more than 3,500 classes. For a more detailed overview of the facilities in the FCL, see Chapter 5.