Gl_itch^es as a generative design process

Glitches - unexpected coding errors - are often discarded as they are considered as design failures. This is surprising as they relentlessly flourish on the web, due the interest their aesthetic triggers. In this paper, we report a preliminary work that aims at understanding glitches in the context of information visualization. We first conducted an empirical study by collecting glitches examples on social media that we grouped by visual and semantic similarities. Glitches descriptions by their creators allowed us to grasp the interest behind sharing such a failure. However, understanding the (unattended) cause of glitches remains usually difficult as it also remains unknown for their creators; nonetheless we reverse engineered some of them in a synthetic manner. We discuss glitches expressive possibilities, which can be considered as a generative design method for visualization, and suggest including them in design studies as a way to better document design processes.