This article argues for the development of a framework through which to describe children's multimodal texts. Such a shared discourse should be capable of including different modes and media and the ways in which children integrate and combine them for their own meaning-making purposes. It should also acknowledge that multimodal texts are not always or only screen-based. In addition, it is argued that current definitions of literacy do not readily answer to the variety of semiotic resources deployed in the design of multimodal texts. In revisiting the author's previous tentative thoughts about "the rhetoric of design" the article develops this theme further through offering a possible framework and using this to analyse three different types of multimodal texts created by seven-year-old children. The framework is, however, a "work in progress", which it is hoped, will open up debate. (Contains 3 tables, 2 figures and 1 note.)