I’ve had some interesting discussions by email since posting my article about software engineering and industrial design. A couple of them have asked where they could learn more, but unfortunately, I don’t have much to offer. King and Chang’s Understanding Industrial Design has a promising title, but telling programmers to make their products simple, enduring, playful, thoughtful, and beautiful isn’t actually very useful: nobody would suggest doing the opposite, and worshipful paeans to classic products doesn’t give readers the intellectual affordances they need to figure out what to do or not to do when designing things themselves.

What would be more useful would be something like Jeff Johnson’s classic GUI Bloopers, which shows readers specific mistakes in specific interfaces and how to fix them. It’s my favorite book on design because it doesn’t waste time trying to convey principles that inevitably sound banal when written down. Instead, its carefully-chosen examples encourage readers to infer those principles for themselves. “I know it when I see it” is a rather imprecise learning objective, but a useful learning outcome.

The other result of these email discussions is the realization that it’s time to re-think how we think about data visualization. Tufte’s classic books (which are really the same book written three times) draw on graphic design, but given how interactive today’s visualizations can and should be, I think that industrial design would be more fertile soil to till. This introduction to machine learning and this introduction to numerical optimization are both more powerful because of the interaction they allow, and that interaction is, I think, best framed in the language of industrial design.

So, at the latest in my long series of get-rich-quick schemes for other people, I think a book (interactive, of course) that explored poor design choices in interactive data visualization and how to improve them could have a lot of impact. I also think it would be fun to create. Any takers?