I'm looking into refactoring a 14000 line legacy c# code behind file!! and I know it has quite a few areas of very similar code. Somebody has come along and copied large lumps of code and modified a small % of them (5% or less)!
It would be a great feature to have Resharper be able to identify areas of code that are similar enough and large enough (both factors would have to be tuneable I guess) to consider parameterizing their differences. obviously it would have to analyse the code structure itself rather than just the text of the code (could be different line formatting/ comments/variable names!) even though the code was basically the same.
I suspect that this would not be that difficult for you to do, although actually performing the parameterisation might be a different matter entirely!