[armish]

ICYMI, iGEM organization's overall goal and the projects that are carried out as part of the annual competitions are highly related to this topic: http://igem.org/Main_Page

In general, it is hard to find a biological system that will give you a reliable and generic compiler framework (a la LVM) unless you better scope your problem and/or goal; but there are small problems for which we have nice implementations/solutions (a la software design patterns). The challenging part is to mix and match parts that will help you solve specific problems.

So the answer really depends on what you want to do with such a system and what parts of the already available solutions that you don't like.