We've been hard at work up here on a little prototype for you guys to try.  This attachment is a basic automatic indexing solution built on top of SQL Server 2005 using the Missing Index DMVs that we discussed in our previous blog post:

http://blogs.msdn.com/queryoptteam/archive/2006/04/06/570176.aspx

Effectively, this will periodically determine top index candidates for your workload.  It currently runs in a recommendation mode, but you can also have it run fully automated if you uncomment a line in the file.

We'd like to ask you guys to give it a try and let us know what you think.  We're interested in learing about the kinds of workloads that it can handle and the kinds of workloads that it can't.  We've played with this idea on a few workloads that we have, but we'd like to use this forum to experiment with how we get feedback from our customers to better serve you in the future.

It's also just cool, so give it a try ;).

Legal Mumbo-Jumbo - these are scripts, and you can do with them as you please, but please don't blame us if it breaks your stuff.  It should work alright, but just be aware that we can't support these scripts to the level that we do for our regular product.  We may be able to ship a similar feature in a future release based on the feedback we recieve.

Enjoy!

Jianjun and Conor