From our experiences developing and evaluating a wide range of visual language environments, and those of others, we have found that visual languages often have a major problem with lack of screen real estate. Multiple views are often used to solve the problem by allowing large systems to be modelled by breaking them into smaller parts. However this approach has limitations and thus we have begun to explore a complementary approach - "Zoomable User Interfaces", or ZUIs. We have added some prototype ZUI facilities to Pounamu, a meta-tool we have been developing for building multiple view visual language environments. We motivate the need for ZUIs in visual language tools, illustrate our ZUI extensions to Pounamu, and discuss our experiences to date in building and using these facilities.