Week 20: another week on Dundry. This week focused on a second tier of interactions and visualisations, more complex than last week – but also some of the key differentiators of this product; the things that make it interesting and unusual.

By and large, this was pretty successful. The sensations we’d hope the interactions would generate really seem to be coming to life, and there’s already a clear leaping-off point for Week 21.

The only point that felt like a misstep was early in the week. In preparation for a new feature, I looked at refactoring the graphics rendering: ultimately, changing the SVG elements output, and how the code that generates them works.

Initially, this seemed like a smart idea. By the end of that day, I changed my mind: I’d overhauled a lot of code without anything moving forward, and we weren’t back at the point I’d been at the end of Week 18. So I made a decision. My job on Dundry isn’t engineering. This is an interaction prototype; design rendered as software. To that end, I’d leave the renderer as it is, focus on new functionality, and deliver the new feature in a slightly altered manner – that wouldn’t require such huge refactoring.

And, at the same time: I’d learned something more about the demands of the SVG pipeline that would have to be accused in production – something to go into the documentation, and based on real experience.

The rest of the week went well, and by Friday, a new set of features and interactions were in good health.

Tuesday also saw a brief detour to Cambridge, to give a lunchtime talk as part of Culture Hack East – sharing some of the process and outputs of *Spirits Melted Into Air* to arts and technology professionals.