I've just finished reading this post on Mary Jo Foley's blog, where she signals an idea that comes from Stephen Walli: releasing SQL Server as an open source project.
Walli was a Microsoft employee, so I think he can well understand what type of product SQL Server is for the company. When I see that he thinks that Microsoft needs to rethink its open-source posturing in order to stay competitive and then that one steps on this direction could be releasing SQL Server as an open source project, I'm shocked.
SQL Server is a database server for the enterprise business, the core of every datacenter centered on Microsoft technologies, the repository of tons of vital informations for a company. Why open sourcing a product like that? What's the real added value for the customers? I don't think it exists...
If I see this from a developer perspective, having the sources of SQL Server could be extremely fascinating, but that's not all. We have to think also from the business and customer's perspective and from this point of view there's no real advantages on such decision.
And then, what about the Microsoft's business? Don't forget that Microsoft is a company that makes business, it's not a voluntary group of hobbiests.
SQL Server is a vital part of Microsoft's business (it comes from many years of tremendous works), so why Microsoft should be so stupid to give the code to the community? Maybe to have a SQL Server clone? Or to help other projects (such as MySQL) to grow?
I think that everyone of us, if he could be on the BillG's desk for taking this decision, will answer NO. And there's another aspect that I'm sure it will be true: if SQL Server will be released as an open source project, many customers will abandon it soon (and Oracle will be happy... 
).