The Malleable Glyph (Challenge)

TL;DR

The paper introduces the Malleable Glyph, a new visualization problem and public challenge that encodes scalar magnitude in a static 1x1 inch glyph, evaluated by pairwise comparisons to determine which glyph has higher or lower $x$ or whether they are equal; observers ask 'how much?' under the Illiteracy Rule. It presents a formal evaluation protocol that uses adaptive testing with $d_t = \gamma^t d_0$, computes $A(G,t)$, and derives glyph resolution $R(G) = 2^{\\mathcal{A}(G)}$, with $D(G) = 100.0 / R(G)$, and discusses calibration via the Weber-Fechner law. It also introduces a Python mglyph library, an exchange format for standardized glyph data, and an online self-evaluation tool, inviting the community to contribute diverse designs and validate them under the challenge. The work aims to stimulate cross-disciplinary creativity while maintaining objective comparability, and contemplates future directions in aesthetics, rotation invariance, composite glyphs, and machine learning evaluation of perception.

Abstract

Malleable Glyph is a new visualization problem and a public challenge. It originated from UX research (namely from research on card sorting UX), but its applications can be diverse (UI, gaming, information presentation, maps, and others). Its essence is: carrying as much information in a defined planar and static area as possible. The information should allow human observers to evaluate a pair of glyphs into three possible sortings: the first is "greater", or the second is "greater", or both are equal. The glyphs should adhere to the Illiteracy Rule, in other words, the observer should ask themselves the question "how much?" rather than "how many?". This article motivates the technique, explains its details, and presents the public challenge, including the evaluation protocol. The article aims to call for ideas from other visualization and graphics researchers and practitioners and to invite everyone to participate in the challenge and, by doing so, move scientific knowledge forward.

Figures (13)

• Figure 1: Illustration of the Q-methodology: subject sorts a deck of cards with subjective utterances into a predefined area typically shaped according to normal distribution. (illustration credit: Adriana Buchmei)
• Figure 2: Card decks for UX testing of the card sorting user interfaces: left: cards with textual cardinals, right: cards with a scaled symbol.
• Figure 3: Preliminary testing showed that different scaled shapes (line, square, circle, star, character shape) have varying degrees to which a human subject recognizes their sizes.
• Figure 4: Malleable Glyphs do not have to be limited to using simple geometric shapes and scaling them. They can use any visual elements: raster graphics, fractal geometry, colors and color scales, textures, clouds of simple discrete elements, as well as anything else. (image credit: https://stockcake.com/i/crystalline-winter-magic_1511734_1170668, https://commons.wikimedia.org/wiki/File:Mandel_zoom_00_mandelbrot_set.jpg)
• Figure 5: Evaluation UI central element: the human subject is given two glyphs of the same kind with different or equal parameters $x$ and makes a decision.