Reflections on metaprogramming

The spread of object-oriented technology has led to object-oriented programming languages with object-oriented implementations. By encapsulating part of t h e semantics of a language within a set of default classes and empowering the programmer to derive new versions of these base classes, a designer can provide a language whose semantics can be tailored by individual programmers. The degree to which such languages are simultaneously flexible and efficient is an open question. We address this question by reporting our experience with using this technique to incorporate support for persistence into the Common Lisp Object System via its metaobject protocol. For many aspects of our implementation we found that the metaobject protocol was perfectly suitable. In other cases we had to variously extend the protocol, pay an unacceptable performance penalty, or modify the language implementation directly. Based on our experience we propose some improvements to the protocol. We also present some performance measurements that reveal the need for improved language implementation techniques.