I've been hearing a lot about PowerShell recently, and was wondering whether there's any reason to use PowerShell instead of (for example) a console application or Windows service using WMI behind the scenes.
What are the benefits and advantages of PowerShell? What is it for?
PowerShell can do lots of things that a .NET console application can do, but it's still a scripting language. But it's a heck of a lot better than the batch file language or VBScript/JScript. It's strengths include being able to use UNIX-like pipes and filters but with objects instead of dumb text. It can also create CLR and COM objects and invoke their properties and methods.
However, if you are writing anything complex, you'd be better served writing a compiled program so you have the benefits of finding errors at compile time and having a better IDE. Also, if you want to run a PowerShell script, PowerShell must be installed on the computer where you're running it.
PowerShell is an interactive shell like KornShell, Bash and er CMD.exe. And like those shells it supports a scripting language (KSH, Bash, Batch). However PowerShell is built on top of .NET and exposes .NET types and allows you to create and manipulate many .NET types. So you can use PowerShell to build scripts that can do what a typical .NET console app could do.
One factor to consider when writing little utilty console applications is how much effort you spend writing parsing & usage code versus the code required to achieve the fundamental purpose of the exe. I have switched to writing a lot of utilities as PowerShell scripts because PowerShell provides a parameter parsing engine with lots of nice features: named/positional parameters, required/optional parameters, default values for parameters, partial specification of parameter names, etc.
PowerShell 2.0 adds even more features in this area (validation attributes, etc) with advanced functions. You can easily write "man pages" as comments for your scripts or advanced functions. Whereas I used to spend 50-80% of my time messing with flaky, non-standard (is it - or / or both?) parameter parsing code in a C# console app, I let PowerShell handle that for me. However I do believe Jacob is correct in saying that complex tasks requiring lower-level .NET code would be easier to get right (static compile time checks) and debug in C#/VB/Visual Studio.
I would love to see the PowerShell parameter parsing functionality exposed via a set of .NET types in the BCL such that you could write a console app and get PowerShell's parsing functionality. A long time ago I used an open source component called Genghis but I think it has been abadoned. At one point during the .NET 4.0 betas, a command line parser appeared in the framework but was removed before RTM. I don't think this command-line parser had any connection to PowerShell - and it should have IMO. So probably a good thing it was pulled.
Powershell's strength is the simplicity and interopterability specifically with Microsoft products. For example, inside Exchange 2007 and 2010, there is no way to modify some attributes without using Powershell. Not until SP2, could you even manage Public Folders in Exchange 2007.
Secondly, I absolutely love, how everything in PoSh is an object. Things like reading in a CSV file are as simple as
$csv = import-csv c:\MyUsers.csv
You now have an indexed object, with each column accessible by it's column head. Not something easily accomplished in many other languages, afaik.
$csv[1].UserName