Windows Server 2008 R2 : Automating Tasks Using PowerShell Scripting - Understanding Shells

Both CLI and GUI shells have benefits and drawbacks. For example, most CLI shells allow powerful command chaining (using commands that feed their output into other commands for further processing; this is commonly referred to as the pipeline). GUI shells, however, require commands to be completely self-contained. Furthermore, most GUI shells are easy to navigate, whereas CLI shells require a preexisting knowledge of the system to avoid attempting several commands to discern the location and direction to head in when completing an automation task. Therefore, choosing which shell to use depends on your comfort level and what’s best suited to perform the task at hand.

A Short History of Shells

The first shell in wide use was the Bourne shell, the standard user interface for the UNIX operating system; UNIX systems still require it for booting. This robust shell provided pipelines and conditional and recursive command execution. It was developed by C programmers for C programmers.

Oddly, however, despite being written by and for C programmers, the Bourne shell didn’t have a C-like coding style. This lack of similarity to the C language drove the invention of the C shell, which introduced more C-like programming structures. While the C shell inventors were building a better mousetrap, they decided to add command-line editing and command aliasing (defining command shortcuts), which eased the bane of every UNIX user’s existence: typing. The less a UNIX user has to type to get results, the better.

Although most UNIX users liked the C shell, learning a completely new shell was a challenge for some. So, the Korn shell was invented, which added a number of the C shell features to the Bourne shell. Because the Korn shell is a commercially licensed product, the open source software movement needed a shell for Linux and FreeBSD. The collaborative result was the Bourne Again shell, or Bash, invented by the Free Software Foundation.

Throughout the evolution of UNIX and the birth of Linux and FreeBSD, other operating systems were introduced along with their own shells. Digital Equipment Corporation (DEC) introduced Virtual Memory System (VMS) to compete with UNIX on its VAX systems. VMS had a shell called Digital Command Language (DCL) with a verbose syntax, unlike that of its UNIX counterparts. Also, unlike its UNIX counterparts, it wasn’t case sensitive, nor did it provide pipelines.

Somewhere along the way, the PC was born. IBM took the PC to the business market, and Apple rebranded roughly the same hardware technology and focused on consumers. Microsoft made DOS run on the IBM PC, acting as both kernel and shell and including some features of other shells. (The pipeline syntax was inspired by UNIX shells.)

Following DOS was Windows, which went from application to operating system quickly. Windows introduced a GUI shell, which has become the basis for Microsoft shells ever since. Unfortunately, GUI shells are notoriously difficult to script, so Windows provided a DOSShell-like environment. It was improved with a new executable, cmd.exe instead of command.com, and a more robust set of command-line editing features. Regrettably, this change also meant that shell scripts in Windows had to be written in the DOSShell syntax for collecting and executing command groupings.

Over time, Microsoft realized its folly and decided systems administrators should have better ways to manage Windows systems. Windows Script Host (WSH) was introduced in Windows 98, providing a native scripting solution with access to the underpinnings of Windows. It was a library that allowed scripting languages to use Windows in a powerful and efficient manner. WSH is not its own language, however, so a WSH-compliant scripting language was required to take advantage of it, such as JScript, VBScript, Perl, Python, Kixstart, or Object REXX. Some of these languages are quite powerful in performing complex processing, so WSH seemed like a blessing to Windows systems administrators.

However, the rejoicing was short-lived because there was no guarantee that the WSH-compliant scripting language you chose would be readily available or a viable option for everyone. The lack of a standard language and environment for writing scripts made it difficult for users and administrators to incorporate automation by using WSH. The only way to be sure the scripting language or WSH version would be compatible on the system being managed was to use a native scripting language, which meant using DOSShell and enduring the problems that accompanied it. In addition, WSH opened a large attack vector for malicious code to run on Windows systems. This vulnerability gave rise to a stream of viruses, worms, and other malicious programs that have wreaked havoc on computer systems, thanks to WSH’s focus on automation without user intervention.

The end result was that systems administrators viewed WSH as both a blessing and a curse. Although WSH presented a good object model and access to a number of automation interfaces, it wasn’t a shell. It required using Wscript.exe and Cscript.exe, scripts had to be written in a compatible scripting language, and its attack vulnerabilities posed a security challenge. Clearly, a different approach was needed for systems management; over time, Microsoft reached the same conclusion.