Brad has coined the term 'Atomic design' (although there has been a slight and friendly banter with Andy Clarke about who first mentioned this as an idea how elements of a design can be organized) for his idea for a methodology to create the 'lego' bricks for a (web) design system.

I think the real power in his idea is not that it is new (it is not), but that the analogy to atoms, molecules, organisms, and the assembly into templates/pages very much encompasses the endless variety that such a 'simple' system can produce, which is exactly what designing for the web is all about. Small bits glued together, but if done right, able to be rearranged to react to environmental changes.

When I first heard Brad talk about this idea - I think it was at a Beyond Tellerand Conference in Dusseldorf (or maybe a early Smashing Conference in Freiburg?) - I rembered how I tried to 'sell' the idea of "designing systems, not 'screens'" to graphic designers I needed to work with: Think about your CI manual for a brand. Your logo treatment on business cards, brochures, trucks and signs, your typographic rules for a variety of use cases across different publications etc... and now think of your 'webdesign' in the same way, design the different parts, and then make a layout. This, and/or the Lego brick analogy, can really help to get away from the 'screen' as the design deliverable.

But Brad's Atomic Design is not only an idea, he has designed a toolset, pattern lab, around this idea, which can help to create a living and maintainable design system.

Atomic Design details all that goes into creating and maintaining robust design systems, allowing you to roll out higher quality, more consistent UIs faster than ever before. This book introduces a methodology for thinking of our UIs as thoughtful hierarchies, discusses the qualities of effective pattern libraries, and showcases techniques to transform your team's design and development workflow.

I'm not sold on all parts of the idea, but I think that's down to implementation details -- since I work/design in parallel between front and back end, usually I won't need to simulate 'real' data in the templates. But that's a luxury very specific to my working/size of projects, and I'm not sure if it'll scale :-)

Status: Read