Poll Results: IoC containers for .NET
Poll Results: IoC containers for .NET
Here are the results for the IoC containers for .NET poll. If you had a look at this page earlier and it looked different, don’t worry – I changed it because somebody had started abusing the poll, so I had to stop accepting submissions. I analyzed the data and created some nice charts.
Here’s the first one: how many people actually use IoC vs. not.
Then, which IoC containers are being used?
And then I’ve got the results per container and per possible rating. If you were taking part in the poll, you may remember that there was a bit of trouble because the Google polling system didn’t allow for all the radio buttons everybody would have liked to see etc… for the results I have now integrated the one important missing item ("heard of it but never tried") into the same overviews with all the other values.
The summary graphics are a bit too large to look nice in-place, so just click on either of these two small graphics to see the large summaries.
I’m thinking I might sit down and do a bit of an overall “ranking", by doing some calculations with the various numbers etc etc… well, maybe if I find some time.
Thanks, everybody, for taking part!
9 comments
Guys, it might be very usefull if all negative commenters would tell, what, in detail is "horrible" and therefore could be improved or corrected. 
Ah come on Peter :-) You're at 108 "never heard of it" and only 2 "it's horrible"!Sorry, just joking - Peter is the author of LightCore.
The automatic result overview that Google creates is unfortunately pretty static. I'm planning to publish my own summary of important points in the future, and I'll also look at all the comments, which aren't really shown above at all. Maybe there's some info in there - I want to let the poll run for a while longer before I do this though.
At last there's a comparison for the different IOC frameworks available. I was searching for this for a long time.Great idea and thank you. :-)
I could be wrong, but something seems a little off with these results. This morning, Unity, StructureMap, and Castle Windsor had a comfortable lead, now suddenly all these projects in the bottom half surged way, way ahead. The same projects with the most 'never heard of it' votes. Puzzle.Net, for example, is now double the nearest competitor. Yet, the Puzzle.Net home on SourceForge shows ratings from a whole 2 users (1 up 1 down), and no activity on the project at all for 1565 days. I also found puzzleframework.com, with no posts/updates since early 2008. Can someone enlighten me as to why it is so much better than the rest? If it's not a dead project, is there another home for it?
For me, I'm a Ninject man, myself. Mainly because it works so smoothly with ASP.NET MVC (via Ninject.Web.Mvc), which is my main platform. It takes me only minutes to set up my IoC in my MVC projects.
I'm surprised that unity came out on top, I would think windsor would have dominated.Too bad you had to take the poll down I'd love to have contributed. I personally like Unity, simple, gets the job done and easy to convince others to try. The MS stamp of approval and documentation makes it much easier to sell teams on.
Windsor is just over the top for me, I appreciate its power, but just too much and too little documentation. My experience trying to convince traditional, stick in the mud, devs to try it: Impossible.
My current project uses an old version of Spring.Net, gosh I want to rip it out. Windsor, Unity, StructureMap, Ninject would all be much better.
I'll give LightCore a go.
Hi Seamus,I think it's a bit hard to judge which tool wins. I'm still planning to publish the results of a slightly more complicated analysis, where each product gets points for being mentioned positively and deductions for being mentioned negatively. In that evaluation, even when the parameters are being modified, StructureMap comes out first, followed by Castle Windsor, and Unity is only third.
In the end, trying yourself and making up your mind about your requirements and priorities is unavoidable. Yes, it's a pity I had to take down the poll, but in the end 600+ votes are quite a good number for an overall impression of the majority preferences.
As someone who hasn't used an IoC container this certainly seems like a nice little comparison. A few percentage based comparisons alongside the straight numbers could add to the value though. eg, it looks like 255 people have tried structure map, that gives 62% saying it's great and only 4% disliking it compared to 256 people who've tried unity coming out at only 41% rating it but 15% thinking it's horrible. That sort of thing certainly helps to point me in the direction of what I should try. 
s someone who hasn't used an IoC container this certainly seems like a nice little comparison. 
very Interesting! 
Comment feed for this post
