Thank you very much for your reply!
I totally agree with you: we engineers anyhow must verify our software again on our own. And this is, what I actually already do to some extent: verifying lots of the results coming from Math.NET with acceptance tests in BDD style, which are validated with R. But, this is just "validation by example" lets say - and still no proof of the correctness.
Our customers do scientific research based on the results, our software is providing. For them, it would be quite important to have a solid base, they take the actual values from. Publishing papers with wrong or incorrect data might not be that good for the reputation.
For me, the question remains, if the community here plans to validate the framework against well accepted math libraries, used in scientific research, such as R or SPSS (maybe also Matlab).
Are there plans to do something like this?