I wouldn’t worry about putting up rants or what feedback you get. I’ve read a lot worse. On your 3rd and 4th points, I agree completely.

This may be able to help you:

http://www.lostechies.com/blogs/marcus_bratton/archive/2009/12/06/the-siege-project-siege-servicelocation.aspx

One syntax to learn, lets your swap out one ioc over another in 1 line of code each (plus adds additional functionality you may find useful). My company uses it in production. I’m aware of other groups using it as well.

I wrote a really naive and cheesy ORM (or, what I thought was an ORM) framework in 2003. A friend of mine with a lot more experience I had at the time said to me that writing an ORM is a pretty typical rite of passage for OO noobs. I’d considered myself a pretty savvy OO developer at the time I was pretty offended. Over the next couple of year I learned that I actually knew very little about OO (about as much as your average .NET developer).

That rite of passage has shifted from writing an ORM framework to writing an IoC framework. With all these dudes trying their hands at writing IoC tools, we’re bound to have a glut of them just as we had a glut of ORM tools.

I probably would have been better off wishing there was only one DI tool in the .NET space instead of asking for one from Microsoft. I definately would have gotten fewer comments.

Maybe I still just don’t have enough knowledge about IOC containers in general, but I didn’t have this problem when picking a unit testing or an ORM framework. There were options, but there were also clearly defined advantages and disadvantages and stated goals for each project. I came out of those comparisons with a clear decision and the feeling like I knew more then I did when I started.

You’ve touched on a persistent core weakness in alt.net: an unwillingness to drive for more understanding of usability in API design coupled with an unwillingness to teach. In this, alt.net continues to pursue ever more insularity which, in my observations, is not exactly undesirable for some of the alt.net folks who are driven more by a selfish desire to be big fish in a small pond.

I think it’s great that you’d step up and voice your frustration. This community needs it’s feet held to a hotter fire and I admire the courage to stick your neck out and call attention to it.

Also, I’d be careful about wishing for an IoC framework from Microsoft. As an organization, they have one of the worst records in understanding the design principles underlying these tools. The last thing the world needs is another Microsoft framework that occults better, simpler, clearer frameworks that it failed to understand.

> I’m very sorry that having options disturbs you so much and I’m very sure you’ll be happy using MEF OOTB with VS 2010 no matter how hard its usability blows.

Actually, poor usability would have allowed me to discard a choice pretty quickly.

> Truth be told, if all you’re going to use is simplistic usages of an IoC tool, any of them work just fine.

Making the entire day I spent cafefully weighing my options time well spent. Sigh.

I am not asking for a Microsoft built IOC container specifically. Yesterday, I was asking for one built into the framework.

I am currently comparing 6 different open source IOC containers trying to make an informed decision. My rant wasn’t anti-open source, it was frustration at not having a clear path forward.

If it was in the framework, I would have spent my day learning the version in the framework, instead of spending the day comparing half a dozen different tools trying to make an informed decision.

If the version in the framework didn’t meet my needs, then I could go looking for another option, armed with a specific need instead of a set of partially understood requirements.