Implementing FS0 in Isabelle: Adding Structure at the Metalevel

Often the theoretical virtue of simplicity in a theory does not fit with the practical necessities of use. An example of this is Feferman's FS0, a theory of inductive definitions which is very simple, but seems to lack all practical facilities. We present an implementation in the Isabelle generic theorem prover. We show that we can use the facilities available there to provide all the complex structuring facilities we need without compromising the simplicity of the original theory. The result is a thoroughly practical implementation. We further argue that it is unlikely that a custom implementation would be as effective.