Modelling has made significant strides since the introduction of the UML in the 1990s and now plays a role in most software engineering projects. However, it is still usually regarded as a means-to-an-end rather than an end itself, with "code" (in third generation, high level programming languages like Java) still being regarded as the primary software artefact driving software projects, especially those following "agile" development methods. The question addressed by this panel is when, if ever, will this situation change in mainstream software engineering practice and when will models taken centre stage in software projects rather than play a supporting role? In other words when will high level, 3GL "code" become irrelevant to software engineers, just as assembler code, machine code and byte code are irrelevant to them today. Another way of casting the question is "when will models become the code?" This will be the day when modelling finally fulfils its promise of raising the level of abstraction at which software systems are described, and high-level source code will simply become an internal, machine-oriented representation format.