I'm designing a thick UI diagnostic tool, should it have a direct integration to PowerShell

Question

My team and I are designing a diagnostic test tool as part of our next product. The test tool will exercise a request/response API and display asynchronous events. As part of the diagnostic tool suite, we will also be providing cmdlets for the entire product API.

Is it worth embedding PowerShell execution into the tool UI ? What are other development teams doing ?

The scripts can still run stand alone in any PowerShell window or tool. From a user's perspective, they would gain the ability to launch scripts from our UI. And, since the UI can be monitoring the same devices that the scripts act on, it brings some unity to the execution of a script and monitoring of the results. Embedding script execution brings more work to the project and I'm not sure how we want to handle displaying the results of the scripts.

Do most PowerShell users expect to run their scripts from their own shell environments or within tools that come from their product vendors ? Note, our diagnostic tool will not be automatically generating scripts for the users as some Microsoft tools do (it might be valuable for inexperienced PowerShell users, but we are expecting most scripts to be fairly simple, like executing a command on a series of devices).

Answer 1

Fortunately embedding the PowerShell engine and execute commands/scripts and getting the results back is pretty trivial. That said, I'm not sure you scenario is one where I would embed PowerShell. You ask if folks prefer to run scripts from their own shells or from within the Tool Vendors environment. I can't speak for everybody but the shells and editors that I use support some nifty features for debugging, code folding, syntax highlighting, multiple runspaces, etc. I'm not sure you would want to go through the effort to provide similar capabilities.

One reason to embed PowerShell is to execute the same PowerShell cmdlets as part of your core diagnostics and monitoring engine. That way you don't have to duplicate functionality between your diagnostic tool app engine and the cmdlets that your customers use for automation. It sounds like the code you use to do the diagnostics and monitoring in the app is different than the code in the cmdlets? Or is there common code shared between the app and the cmdlets?

Another reason to embed PowerShell is to allow the app itself to be scriptable but this doesn't appear to fit your scenario.

Another reason to embed PowerShell is if you are implementing a new host - ie you provide some unique editing or shell functionality. Some apps that do this are PowerGUI (which allows you to launch scripts IIRC) and PowerShell Plus.

Yet another reason I have embedded PowerShell in an application is because I knew I could get certain results in much less code than the equivalent C# code. This is a weaker reason and I probably wouldn't do this in a commercial app but I have used this for one-off programs.

Answer 2

I agree with both Jaykul and Keith Hill - the answer is yes.

There are several approaches you could use. But in general, I'd recommend you a) create key cmdlets as part of the UI for your app and b) you build the GUI on top of PowerShell (in the same way the Exchange team has done.

Doing this follows Microsoft's lead (all applications have to have a PowerShell interface) that is also being taken up by others (e.g. VMware, and even Symantec leverage PowerShell in their applications.

Creating cmdlets (and possibly a provider) is pretty straightforward - there's a great cmdlet designer recently released (see http://blogs.msdn.com/powershell/archive/2009/10/16/announcing-open-source-powershell-cmdlet-and-help-designer.aspx) for this tool.

Hope this helps!