The considerations I've done yesterday on when using SQL CLR features against T-SQL must be read with the right sense. What are the keys to decide for a technology or to take a project decision?
Obviously, performance is one of the most important key when you've to decide the technology to use and user's satisfaction comes as a consequence. But this is not all.
When building applications for business there are other factors that must be take into considerations, such as the scalability of the solution (if over time the application's load is growing, are you ready?), the maintainability of the solution you're building over time (the code must be easy to maintain and upgraded), the direction of the future technology (the solution must be created with respect of the future standards and must be rapidly adaptive to future technologies), but also productivity (you've to decide the right solution also with taking into considerations how much time you've to spent on the project).
And this last aspect is what my previous post about SQL CLR wants to take into considerations: if you've to spent lots of days of work to build a complex routine with T-SQL, maybe difficult to maintain in the future, take into considerations the .NET way: your routine could be deployed with a .NET language and its maintainability in the future could me more and more easy. 