The search for a type-safe version of Eiffel is a never-ending story in the development of the Eiffel language. Recently, Helmuth Brandl has created a series of blog entries with a nice description of is needed to make Eiffel type-safe. There have been others posting blog entries (including me) and it also had been a constant topic during the ECMA process of Eiffel standardization.
Interestingly, a question that was seldom raised was: why do we need type-safety? As absurd as it seems for all the mathematicians to raise this question, the Eiffel community has understood for a long time that mathematical soundness is not a reason in itself. Instead, it serves a purpose in the enterprise of language design: giving software developers means to express themselves in ways that reduce the number of defects in his/her programs and enable him/her to master complex problems he/she would not be able to master without.
There are many discussions on how to make Eiffel type-safe. The ECMA committee has been very active in coming up with solutions (though non of these solutions have been fully implemented and analysed).