Octavio Pardo Virto After more than 200 years of research, the topic of legibility still seems slippery when it comes to understanding the impact of typographic factors. The practical use of research papers for type designers sometimes seems unclear. Complex font variables like weight, width, or contrast are dealt with as if they were binary parameters instead of the continuums they represent in the type design process.
In this presentation. I will show the results of my Ph.D., where I have developed a new experimental paradigm using Bayesian Algorithms (AI) and a variable font. The experiment consists of a reading speed test for 22 participants, with 100 trials. The AI algorithm explores the tridimensional space of the variable font seeking to find those font configurations that increase the reading speed for each participant. The output of the experiment are font coordinates that can be directly transformed into new typefaces by type designers. This approach turns my research into a conversation and simplifies font research making it a collaborative project.
AI and variable fonts together might ignite a new era for font customization, legibility, and low-vision solutions. However, despite the amazing range of possibilities that this thesis presents for type designers, it also raises a concerning question for them. Is legibility still a font issue or has the debate switched to a technology issue?