Should I bother learning Windows Workflow?

Question

Okay, so I've learned WPF and WCF and that was great. I didn't bother with CardSpace because it's probably not gonna take off.

Should i bother with Workflow,..is it worth it? What's the big deal?

Answer 1

It depends on the type of software you build. WPF is neat. WCF has its uses. Windows Workflow is pretty "enterprisey", though. It requires more setup and can get complex.

Unless you build enterprise software that would benefit from a definable workflow, I wouldn't bother much with it.

Answer 2

Windows Workflow is useful for several different kinds of application. If you're not writing one of them, then there's no need to bother learning it.

1. Human workflows - obvious
2. It can be useful for complex UI navigation
3. It can be very useful as a way of processing several operations in parallel. It's a very easy way to do asynchronous processing.
4. Long-running operations consisting of multiple steps
5. Rules-based processing
6. Any situation where you need to permit your customers to customize behavior. You can ship custom activities that pertain to your application, and customers can then combine and parameterize them.

But if you don't need any of this, why bother? Just read some of the whitepapers, or look at some of the videos and see if any of it looks interesting.

One interesting thing that's not obvious: both the workflow designer and the rules designer can be hosted in arbitrary Windows Forms applications. This allows them to be reused in your own application, if, for instance, you need a way to customize conditional processing.

Answer 3

Like Paul said, it depends on the environment you work in. Personally, since I never know where life is going to take me (corporate development, startup, etc.) I try to learn everything I can when I have the chance.

My advice would be, if you have the time, the intuition, and the development environment...learn it. You never know when it might come in handy at some point down the road.

Answer 4

I think the general consensus on the current Windows Workflow is that the product is not quite ready for primetime yet. If you want to get anything more than trivial accomplished with the current framework, you will likely have to write a few custom tasks to fill in functionality gaps. I've also had issues with the designer flaking out when the complexity of the workflow starts piling up.

From what I've been hearing/reading about Windows Workflow, the next version (v4.0) is a complete rewrite that fixes a lot of the issues in the current version. They claim backward compatibility with v3/3.5 workflows, but you probably will have to know the framework pretty well to know what changed and how to fix it if it doesn't work right away.

Unless you have a project that absolutely demands this functionality today, you're probably better off waiting until the new bits ship.

Answer 5

It is worth learning or at least understanding WF. WF offers a different paradigm for building apps, a paradigm which may become mainstream in the future. Enterprisey? Yes, maybe today. But longer term WF or at least the paradigm it promotes, may become ubiquitous and universal. Using Workflows to describe programs may today be in the stage where OO and C++ was, 20 years ago.

Learning a new thing broadens your thinking. For example, even if you use C# or Java in your day-to-day job, knowing Lisp helps broaden your thinking about how to solve problems. IMO. WF is the same.

A second thought is to think like an investor. If you want to raise the value of YOU as a programmer, whether to learn WF would depend on the market or job you are in. WF skills in your space or in your company may or may not be worth your extra effort.

Answer 6

Ever seen Leon Bambric's post on Workflow software?

Answer 7

Thanks for all the advice guys. Cheeso get's the answer since he got the most up votes but all opinions are appreciated. I think I'll take Scroll Lock's advice and wait until v4.0 until I take a serious look.

Answer 8

I originally really liked the concept of workflows, but in the end I think I'll just use something like stateless or roll my own statemachine, and use a database for persistence. My biggest issue with it right now is that there still seems to be a huge memory leak in the workflow runtime. Makes me kind of nervous.