For more than 10 years, I had the honour and pleasure to work with Celia Hoyles and Richard Noss. We share a common concern for more learnable mathematics, especially in algebra, and for the need to build new representational infrastructures taking advantage of technology. Beyond this common concern, my choice to work in the French institutional context and paradigms for the Casyopee project led to a different approach. Roughly speaking, in these context and paradigms the stake is to help students access existing representations while Hoyles and Noss privilege building new more learnable representations. The aim of the article is to investigate empirically how, in spite of this gap, Hoyles and Noss' approach can be an inspiration for enlarging the focus on a given topic, freeing oneself of a too strict dependency to one's own context.