The Dimensions of Perception: Evaluating the Value of Depth, Scale, and Presence in Web-Based Geometric Artworks

Full report

On this page

Introduction to Spatialized Geometric Art

The transition of digital geometric artworks from two-dimensional rasterized images to interactive, web-based three-dimensional environments represents a fundamental shift in human-computer interaction and visual perception. Modern web standards, powered by WebGL, WebAssembly, and the WebXR Device API, have democratized access to complex spatial rendering1. These frameworks allow users to experience intricate geometric topologies directly within the browser, eliminating the need for specialized software installations or walled-garden applications3. However, this technological capability prompts a critical evaluation: at what point do the addition of stereoscopic depth, immersive scale, and spatial presence cease to be mere technological novelties and begin providing tangible cognitive and aesthetic value?

Geometric artworks—particularly those characterized by complex algorithmic generation, fractal scaling, and dense node-link structures—pose unique perceptual challenges. On a standard two-dimensional desktop display, high-density geometric visualizations often suffer from visual occlusion, colloquially known as the "hairball effect," where overlapping lines and shapes destroy topological clarity and severely limit analytical or aesthetic engagement5. The introduction of immersive virtual reality (VR) headsets provides binocular disparity and motion parallax, depth cues that theoretically resolve these physical and visual occlusions7.

Yet, immersive technologies also introduce significant barriers, including cybersickness, vergence-accommodation mismatch, elevated cognitive load, and accessibility constraints9. By comprehensively evaluating empirical literature on graph visualization, spatial memory, user interface legibility constraints, physiological human factors, and the W3C’s Extended Reality Accessibility User Requirements (XAUR), this analysis delineates the specific conditions under which desktop 3D and headset VR deliver measurable value to geometric spatial computing.

1. Navigating Topological Complexity: Cognitive Load and Stereoscopic Depth

The foundational debate in spatial computing revolves around the utility of the third dimension. Early usability heuristics established by the Nielsen Norman Group cautioned against the superfluous use of 3D, noting that navigating three-dimensional computer spaces imposes a high cognitive burden on users interacting with standard user interfaces, advocating instead for the simplicity of 2D surfaces11. However, geometric art and abstract spatial data—such as 3D line sculptures or complex network visualizations—operate under entirely different perceptual paradigms than traditional user interfaces13.

1.1 The Mitigation of the "Hairball Effect" Through Stereopsis

Empirical research consistently demonstrates that interpreting large, complex geometric structures is significantly enhanced by 3D viewing, provided that the appropriate depth cues are present15. When viewing a dense 3D geometric web on a 2D desktop monitor, users rely entirely on monocular depth cues such as perspective, shading, and occlusion7. While useful, these cues fail when geometric density reaches critical thresholds. Research into the scalability of network visualization indicates that physiological measures of cognitive load (measured via electroencephalogram) increase rapidly as geometric complexity grows5. For instance, in small-world graphs featuring 50 or more nodes with a density ratio of six edges per node, participants utilizing 2D desktop displays were unable to correctly answer spatial queries in more than half of the trials5. The usefulness of standard node-link diagrams deteriorates exponentially, resulting in visual "hairballs" that demand significant simplification or filtering to remain intelligible5.

The introduction of immersive headset VR directly addresses this perceptual bottleneck. Ware and Mitchell demonstrated that the combination of kinetic depth cues (motion parallax) and stereoscopic depth cues (binocular disparity) dramatically increases the size and complexity of a visual structure that a human can accurately interpret15. Their research revealed that the use of a head-coupled stereoscopic display—mimicking the spatial tracking of modern VR—reduces error rates in path-tracing tasks by an order of magnitude15. By leveraging both structured 3D motion and stereo viewing, untrained observers could accurately parse networks of over 333 nodes with an error rate of less than 10%, while skilled observers could navigate geometric structures comprising up to 1,000 nodes with similar accuracy15.

In a WebXR context, a geometric artwork that appears as an unintelligible mass of lines on a 2D canvas becomes structurally coherent inside a VR headset because the human visual system's super-acuity is optimized for stereopsis and kinetic movement8. The true value of VR over desktop 3D emerges precisely when the geometric density of the artwork exceeds the natural occlusion limits of monocular projection.

1.2 Stereoscopic Highlighting and Spatial Focus+Context

Beyond raw depth perception, VR allows for novel visual emphasis techniques that are impossible on 2D desktops, preserving the native aesthetics of the geometric artwork. Alper et al. introduced the concept of "stereoscopic highlighting" for spatial visualization8. Rather than relying solely on color adjustments or motion to highlight a specific geometric sub-structure—which may interfere with the artwork's intended aesthetic palette—stereoscopic highlighting brings regions of interest closer to the viewer exclusively along the Z-axis (depth)19.

This methodology creates a spatial "Focus+Context" view19. The highlighted geometric nodes and their immediate topological neighbors are rendered at a closer stereoscopic depth, while the overall macro-structure of the graph is pushed further back into the virtual volume19. In controlled experiments, while stereoscopic highlighting alone did not always drastically outperform static color highlighting for simple tasks, the combination of both techniques yielded significantly faster completion times and higher accuracy in complex topological analysis8. For interactive web-based geometric art, this implies that VR headsets provide developers with a dedicated, invisible Z-axis channel for interactivity and structural emphasis, preserving color and luminance exclusively for aesthetic expression.

1.3 Spherical Layouts and Depth Routing in Immersive Environments

The translation of traditional 2D graph layouts into 3D space often requires specialized routing to maximize the benefits of immersive viewing. When a standard 2D layout is mapped directly onto a 3D sphere, distortion increases the further a point moves from the tangent plane, resulting in nodes becoming compacted8. Kwon et al. demonstrated that employing spherical graph layouts combined with depth routing significantly outperforms traditional flat layouts when viewed in an immersive environment7.

By projecting nodes onto the surface of a sphere and routing edges through the interior depth of the sphere, developers can mitigate the "zero angular resolution problem" (where nodes and edges obscure one another)21. In rigorous testing across datasets ranging from 34 nodes (60° × 33.75° Field of View) to 297 nodes (150° × 84.375° Field of View), participants utilizing stereoscopic depth routing in VR answered complex spatial queries significantly faster and with fewer interactions than those utilizing traditional 2D representations7. For geometric artworks designed to envelop the user, spherical depth routing provides a mathematically rigorous foundation for ensuring that aesthetic complexity does not degrade into visual chaos.

2. Spatial Memory and the 2D vs. 3D Efficacy Paradox

While immersive VR excels at resolving topological occlusion and rendering vast geometric structures, its impact on spatial memory and learning is highly paradoxical. The assumption that providing users with a physically realistic, three-dimensional spatial environment universally improves memory acquisition has been heavily scrutinized in human-computer interaction literature.

2.1 The Deterioration of Spatial Memory in Unconstrained 3D

Cockburn's foundational evaluations of spatial memory across 2D, 2.5D, and 3D physical and virtual environments revealed that the addition of a third dimension does not inherently improve a user's ability to locate and recall items22. In fact, providing users with unconstrained freedom to locate elements in a 3D volume often leads to a deterioration in recall performance. Users frequently perceive interfaces with higher dimensions as more "cluttered" and less efficient, requiring slower sequences of mechanical actions to navigate compared to the rapid, indexed nature of 2D screen interactions22.

This phenomenon is mirrored in educational and operational contexts. In studies evaluating the memorization of complex building components (Building Information Modeling) and pipefitting tasks, results have frequently shown conflicting data regarding the superiority of immersive VR over traditional 2D isometric drawings25. In a study involving biology students utilizing 2D models versus 3D virtual models, the 2D visualizations yielded superior learning outcomes; the researchers theorized that the 3D virtual models created cognitive distraction, drawing mental resources away from spatial encoding and toward environmental navigation25. Therefore, if a web-based geometric artwork is designed strictly as an informational dashboard or a tool for precise, unhindered data acquisition, forcing the user into a 3D VR environment actively degrades their operational efficiency.

2.2 The Method of Loci and Landmark-Based Encoding

The true efficacy of 3D spatial memory relies heavily on the environment's ability to support the "method of loci," a mnemonic strategy where spatial landmarks assist in memory encoding22. In virtual environments populated with prominent background landmarks, spatial presentations result in significantly better memory scores and enable higher-level visualization strategies during memorization tasks27.

Furthermore, behavioral studies evaluating wayfinding within topological immersive virtual environments (TopoIVE) illustrate how users organically divide their reliance between 2D and 3D representations when given the choice28. Participants overwhelmingly preferred 3D modes (spending 85% of their time in 3D) for active movement, aesthetic appreciation, and gaining a realistic sense of scale28. However, these same users explicitly switched to 2D modes for orientation, overcoming visibility problems, and removing the occlusion created by large virtual structures28.

For the developer of web-based geometric art, these findings dictate a hybrid approach. If the artwork is designed as a navigable architectural space where the geometric lattice serves as a structural landmark, immersive VR effectively harnesses episodic and spatial memory27. Conversely, providing a 2D "minimap" or an orthographic toggle remains essential for users who become disoriented within the sublime complexity of the 3D geometry.

3. The Affective Matrix: Presence, Aesthetic Awe, and Therapeutic Utility

Beyond cognitive parsing and spatial memory, geometric art is fundamentally an aesthetic experience. The transition from a bounded desktop monitor to a fully immersive VR environment triggers profound affective responses, primarily categorized under the psychological constructs of "presence" and "aesthetic awe."

3.1 Measuring Spatial Presence and Sensory Prediction

Presence is the subjective sensation of "being there" in a virtual environment, frequently conceptualized as the "place illusion" that persists despite the user's conscious knowledge that they are viewing digital content29. The measurement of this phenomenon is frequently conducted using the MEC Spatial Presence Questionnaire (MEC-SPQ), which evaluates factors such as attention allocation, spatial situation models, self-location, and possible actions30.

Empirical baselines utilizing the MEC-SPQ consistently demonstrate that VR headsets generate significantly higher scores in spatial presence and self-location compared to 2D desktop or mobile interactions30. When users observe a geometric artwork on a desktop, they maintain a clear visual and cognitive separation between their physical room and the digital window. In contrast, immersive VR monopolizes the visual and auditory fields, forcing the cognitive system to build a spatial situation model based entirely on the digital geometry29. Research further indicates a significant correlation between the accuracy of predicting sensory consequences of motor actions (e.g., turning one's head and seeing the geometry update with zero latency) and the perceived realness of the virtual environment29. The artwork ceases to be an object of observation and becomes an environment of inhabitation.

3.2 The Elicitation of Aesthetic Awe and the Impact of Generative AI

Awe is a complex emotion typically triggered by stimuli that are vast—either physically or conceptually—and require a need for cognitive accommodation33. Research into environmental psychology and virtual reality has shown that immersive systems are highly effective at inducing "aesthetic awe," particularly when exposing users to environments like deep space or sublime natural landscapes33.

In geometric artworks, scale is entirely arbitrary. A mathematical fractal can be rendered at the size of a teacup or the size of a galaxy. On a desktop screen, this scale is constantly contextualized and diminished by the physical bezel of the monitor and the surrounding room. In VR, developers can position the user at the microscopic center of a sweeping geometric lattice, inducing a profound sense of vastness33. Exposure to awe-inducing virtual environments is empirically linked to increased physiological arousal, as evidenced by increased fast alpha brainwave activity, indicating a state of relaxed alertness33.

It is worth noting that the perception of such art is also influenced by its perceived origin. Cross-cultural studies investigating machine-generated artworks reveal that when art is explicitly attributed to Artificial Intelligence, users report reduced aesthetic appreciation, even if the visual output is identical to human-created art36. For creators of generative geometric art, framing the narrative of the artwork's creation is as critical to the user's affective response as the rendering technology itself.

3.3 Therapeutic Application: Hypnotic Enhancement of VR Distraction Analgesia

The affective power of immersive geometric environments extends beyond mere entertainment, finding profound utility in clinical settings. VR is increasingly utilized as a non-pharmacologic analgesic to control acute pain during medical procedures37. By monopolizing the user's sensory input, VR acts as a powerful cognitive distraction37.

Recent randomized controlled trials have investigated the additive analgesic efficacy of combining immersive VR distraction with hypnotic analgesia37. In a study involving 205 participants subjected to experimentally induced acute thermal pain, researchers found that being hypnotized during thermal pain significantly enhanced VR distraction analgesia37. While posthypnotic suggestions alone can be limited by an individual's inherent hypnotizability, the combination of a visually overwhelming, awe-inducing virtual environment (such as an infinite, pulsating geometric lattice) with clinical hypnosis drastically reduced the perception of "worst pain"37. This highlights a compelling value proposition for headset VR: the ability of sublime, immersive geometries to override the central nervous system's processing of physical trauma.

4. Web-Based Spatial Architecture: DOM, WebGL, and WebXR

To deliver high-fidelity spatial experiences across diverse hardware—from standard desktop monitors to tethered PC-VR and standalone mixed-reality headsets—developers rely on a sophisticated stack of web technologies. The transition from building 2D Document Object Model (DOM) layouts to rendering 3D spatial environments requires a fundamental shift in architectural philosophy.

4.1 Bridging the UI and GPU: WebGL and WebAssembly

Traditional web pages rely on the browser's layout engine to calculate HTML and CSS, rendering text and elements natively39. Immersive 3D web content, however, abandons the standard DOM compositing pipeline. Instead, it utilizes WebGL (Web Graphics Library), a JavaScript API based on OpenGL ES that executes shader code directly on the computer's Graphics Processing Unit (GPU) via an HTML \<canvas\> element39.

Because WebGL operates outside the standard browser rendering flow, synchronizing DOM elements (like scrolling text) with WebGL graphics presents severe architectural challenges39. To alleviate the steep learning curve of raw WebGL, developers heavily utilize abstraction libraries like Three.js, which handles scene graphs, meshes, and materials4. Furthermore, frameworks like A-Frame, developed initially by the Mozilla VR team, provide an Entity-Component-System (ECS) architecture, allowing developers to author 3D and VR experiences using declarative HTML-like syntax4.

For geometric artworks requiring real-time procedural generation or complex physics simulations, JavaScript's execution speed often becomes a bottleneck. To achieve the predictable frame times required for immersive comfort, developers increasingly compile performance-critical code into WebAssembly (WASM). Benchmark data indicates that complex physics and algorithmic constraints run 5 to 20 times faster in WASM compared to equivalent JavaScript implementations, providing near-native execution efficiency while avoiding garbage collection interruptions that cause visual stutter1. Furthermore, edge computing networks are increasingly utilized to run multiplayer or data-heavy XR computations in data centers physically closer to the user, drastically reducing network latency41.

4.2 The WebXR Device API: Sessions and Reference Spaces

The W3C WebXR Device API serves as the official standard for managing XR device sessions, pose tracking, and controller inputs, replacing the deprecated WebVR API2. The lifecycle of a WebXR experience begins with capability detection, where the application queries navigator.xr.isSessionSupported() to determine if the user's hardware supports the desired mode (e.g., inline, immersive-vr, or immersive-ar)2. Due to the API's strict security model designed to prevent malicious tracking, initiating an immersive session requires explicit user activation (such as clicking an "Enter VR" button) before the navigator.xr.requestSession() promise can resolve42.

Once a session is established, the translation of physical movement into virtual coordinates is governed by the XRReferenceSpace interface35. The selection of the reference space profoundly impacts how the user perceives scale, depth, and mobility within the geometric artwork:

• viewer: The origin tracks the user's head position and orientation exactly35. This is primarily used for inline (desktop) experiences or head-locked UI elements. Overusing this in VR can cause severe cybersickness, as the world appears rigidly attached to the user's face.
• local: Establishes an origin near the viewer's starting position, optimized for seated experiences where the user is not expected to physically translate35.
• local-floor: Similar to local, but the origin's Y-axis is securely anchored to the physical floor44. For geometric art, this is critical. It allows the artwork to establish a 1:1 architectural scale with the user's physical height, ensuring that structures rising from the "ground" possess genuine mass and scale44.
• bounded-floor: Designed for room-scale VR. It provides a boundsGeometry defining the safe physical perimeter in which the user can walk without colliding with real-world objects44.
• unbounded: Allows total freedom of movement over extremely long distances. The viewer is not tracked from a static origin; rather, the origin dynamically drifts to maintain high-precision tracking stability near the user's current location35.

If tracking is lost or a discontinuity occurs between the tracked object's origin and the user's environment, the API fires a reset event on the XRReferenceSpace. Applications must handle this event to recalculate offset transforms (via getOffsetReferenceSpace()), seamlessly recentering the user's view without disrupting the geometric simulation35.

4.3 Frame Timing and Motion-to-Photon Latency

In standard 2D web development, animations utilize the browser's window.requestAnimationFrame. However, immersive environments bypass the standard DOM compositor to minimize latency47. WebXR utilizes a dedicated XRSession.requestAnimationFrame() loop, which operates on the XR device's native refresh rate and provides highly precise pose data for tracking43.

The "motion-to-photon latency"—the total delay between a user's physical head movement and the corresponding visual update on the display—must remain strictly below 20 milliseconds to prevent cybersickness47. Modern mixed reality headsets operate at refresh rates of 90Hz to 120Hz, leaving a rendering budget of just 8 to 11 milliseconds per frame47. Mixing traditional DOM rendering techniques with WebXR render loops creates subtle frame timing issues that manifest as visual judder, instantly destroying presence and inducing nausea47.

5. Spatial Legibility: Typography and the WebXR Layers API

Geometric artworks frequently incorporate textual annotations, mathematical metadata, or interactive UI panels. In a 2D desktop environment, rendering crisp text is natively handled by the browser's layout engine. In WebXR, text must be rasterized into the WebGL texture pipeline, introducing severe legibility challenges due to aliasing, scaling artifacts, and perspective distortion39.

5.1 Distance-Independent Millimeter (DMM) and Vergence-Accommodation

Determining the correct font size in a 3D volume requires accounting for angular resolution. A 16-pixel font on a desktop screen is legible because the user is positioned at a fixed, predictable distance. In VR, a user might view an annotation from 0.5 meters or 15 meters away49. To standardize spatial UI design, Google introduced the Distance-Independent Millimeter (DMM)9.

A DMM is defined as one millimeter in height at a viewing distance of one meter9. The mathematical relationship is linear based on distance: a 41 DMM UI element will physically scale to 82 mm at a 2-meter distance, maintaining a constant angular size on the retina50. Empirical research into VR reading comfort indicates that the optimal text size for prolonged reading in immersive environments is 41 ± 14 DMM49. Furthermore, legibility is heavily impacted by the font typeface and background contrast. Studies demonstrate that users prefer white text on a black background, and sans-serif fonts (like Arial) significantly outperform serif fonts (like Times New Roman) in immersive environments, with two-dimensional flat text performing better than embossed or 3D-extruded text49.

When designing spatial layouts, developers must respect the vergence-accommodation conflict9. Because the lenses in a VR headset focus the eyes at a fixed focal distance (commonly around 1.3 meters), placing UI elements significantly closer or farther than this focal plane forces the user's eyes to converge at one distance while accommodating (focusing) at another, leading to severe ocular fatigue9. To maximize comfort, critical geometric annotations should be placed within the comfortable vergence boundary (0.75m to 3.45m), avoiding the extreme periphery9.

5.2 Signed Distance Fields (SDF) and WebXR Composition Layers

To maintain text and geometric edge sharpness at any arbitrary scale, WebGL developers rely on Signed Distance Fields (SDFs)52. Instead of storing a static pixel grid of a character—which blurs aggressively when scaled up as the user walks closer—an SDF texture stores mathematical distance values from each pixel to the nearest edge of the shape52. The fragment shader uses this distance value to procedurally reconstruct perfectly crisp curves and serifs in real-time54. SDFs are vital for VR geometric art, where users can lean in and view textures at microscopic distances.

However, even with SDFs, rendering text onto the standard WebGL canvas presents a final hurdle. Before the image is sent to the headset's displays, the XR compositor must distort the image to correct for the physical lens curvature (barrel distortion). This post-processing pass degrades the sharpness of the original render, often making small UI text unreadable55.

To combat this, the W3C WebXR Layers API introduces geometric "composition layers" (e.g., XRQuadLayer, XRCylinderLayer, XREquirectLayer)55. Instead of rendering text into the main WebGL scene, developers project the text onto a mathematically defined XRQuadLayer55. The browser bypasses the WebGL canvas entirely and hands this pristine quad directly to the underlying XR compositor (e.g., Oculus OS). The hardware compositor then projects the text at maximum resolution, applying lens distortion perfectly without the aliasing introduced by double-rendering55. Layers can even be assigned layout attributes like text-optimized or graphics-optimized to instruct the compositor on prioritizing readability versus sharpness55. If the text contains critical geometric data, deploying WebXR Composition Layers is indispensable.

6. Physiological Constraints: Reevaluating Cybersickness Assessment

The primary physiological barrier to the widespread adoption of VR over desktop 3D is cybersickness. Cybersickness manifests through symptoms of general discomfort, nausea, disorientation, and oculomotor strain57. It is fundamentally caused by a sensory conflict between the visual system (which perceives aggressive motion through the headset) and the vestibular system in the inner ear (which registers the physical body as stationary)59.

6.1 The SSQ and the Zero-Baseline Fallacy

The severity of cybersickness is universally measured using the Simulator Sickness Questionnaire (SSQ), originally developed by Kennedy et al. to evaluate military flight simulators10. The SSQ asks users to rate 16 symptoms (from "none" to "severe"), outputting weighted scores across three sub-scales: Nausea (N), Oculomotor (O), and Disorientation (D)58.

Historically, VR researchers administered the SSQ exclusively post-exposure, assuming a "zero baseline"—the presumption that healthy users enter the simulation with zero physical symptoms10. However, recent systematic analyses have exposed the erroneous nature of this assumption. Large-scale studies indicate that healthy participants from the general population frequently report non-zero baseline SSQ scores, commonly citing pre-existing fatigue, mild headaches, or baseline eye strain10. In rigorous testing, general populations exhibited a median baseline SSQ score of 8, while healthy sub-populations exhibited a median of 910. Furthermore, factor analysis indicates that certain SSQ items (like general discomfort and difficulty concentrating) inadvertently measure state anxiety rather than genuine cybersickness63.

If a user views a highly dynamic geometric artwork and subsequently reports a moderate SSQ score, failing to account for their baseline physiological state or inherent anxiety may lead a developer to falsely conclude the artwork is nauseogenic10. To address these psychometric limitations, alternative tools like the Cybersickness Questionnaire (CSQ-VR) have been validated. The CSQ-VR demonstrates substantially better internal consistency than the SSQ and benefits from examining pupil size—a physiological biomarker of cybersickness intensity—providing a more accurate assessment of a user's degradation in cognitive and psychomotor skills during VR exposure64.

6.2 Locomotion and Environmental Design

The design of the geometric environment directly dictates the severity of subsequent cybersickness. Studies comparing VR tasks with forced continuous locomotion against stationary observation demonstrate significant variances66. When a user is subjected to linear and angular accelerations—such as flying smoothly through a 3D fractal—the vestibular-visual mismatch maximizes, sharply increasing nausea and disorientation scores60.

Physical hardware factors further exacerbate these symptoms. The weight of the headset (often approaching 800 grams with straps) contributes to increased neck muscle strain and fatigue, which amplifies visual-vestibular conflict during extended immersions61. Conversely, environments utilizing a local-floor reference space—where users physically walk around a stationary geometric sculpture—exhibit significantly lower cybersickness, primarily localized to manageable oculomotor strain rather than debilitating nausea61. For web-based geometric artworks, providing users with the option to observe the art as a stationary object, or utilizing instant teleportation mechanics rather than smooth artificial locomotion, preserves the immersive value of the experience while actively mitigating physiological distress.

7. Accessibility Frameworks and Neurological Safety

As WebXR technologies advance, integrating rigorous accessibility standards is no longer an afterthought but an ethical and functional requirement59. A web-based geometric artwork must comply with standards that allow diverse users, including those with physical, cognitive, auditory, or neurological disabilities, to engage with the spatial content safely59.

7.1 W3C XR Accessibility User Requirements (XAUR)

The W3C's XR Accessibility User Requirements (XAUR) provides a comprehensive framework for inclusivity in immersive spaces, addressing the multifaceted modality requirements of XR59. A core challenge in VR is the reliance on 6-Degrees-of-Freedom (6DoF) spatial controllers, which demand fine motor skills, dexterity, and precise timing, thereby creating insurmountable barriers for individuals with physical or motor disabilities (such as tremors)70.

The XAUR emphasizes several critical User Needs and Requirements (REQ) that must be engineered into the foundation of WebXR experiences:

• Immersive Semantics and Navigation (REQ 1a-1e): Geometric objects and locations must possess robust affordances and accurate semantic descriptions readable by assistive technologies, allowing visually impaired users to navigate spatialized environments via text-to-speech and sonic symbols69.
• Motion-Agnostic Interactions (REQ 2a-2c): The VR environment must not demand specific bodily movements to trigger actions59. If interacting with a geometric node requires reaching to a height of 2 meters or executing a complex spatial gesture, developers must provide a motion-agnostic alternative. This includes supporting multimodality, allowing users to interact via gaze-based triggers, voice commands, or adaptive keyboard input simultaneously69.
• Immersive Personalization (REQ 3b): Users with cognitive or learning disabilities must be permitted to mute non-critical spatial audio, disable aggressive environmental animations, and activate safe harbour controls to accommodate sensory processing sensitivities69.

7.2 Photosensitive Epilepsy and Harding FPA Thresholds

Geometric artworks heavily feature algorithmic pattern generation, strobe lighting, and high-contrast color cycling. Such techniques pose a severe neurological risk to individuals with photosensitive epilepsy (PSE), a condition affecting approximately 1 in 4,000 people, where seizures are triggered by specific visual stimuli73.

To ensure safety, web content must strictly adhere to the Web Content Accessibility Guidelines (WCAG) 2.1, specifically Success Criterion 2.3.1 (Three Flashes or Below Threshold)68. Compliance is frequently evaluated using automated diagnostic tools like the Harding FPA (Photosensitive Epilepsy Analysis Tool), which analyzes video rendering frame-by-frame against established physiological thresholds74.

An immersive geometric sequence will be flagged as a critical violation if it meets the following criteria simultaneously:

1. Temporal Frequency: The content flashes more than three times within any one-second period (\>3Hz)73.

2. Luminance Contrast: The flash involves a peak-to-peak luminance transition of 20 cd/m² or greater, and the darker frame falls below 160 cd/m²74.

3. Spatial Area: The flashing element occupies more than 25% of the central visual field (or \>0.006 steradians)74. Furthermore, high-contrast, oscillating spatial patterns (such as densely packed geometric stripes or checkerboards) with spatial frequencies between 1 and 4 cycles per degree can trigger photoparoxysmal responses even without temporal flashing73.

Because VR headsets monopolize the entirety of the user's field of view, a geometric flash that might only occupy a safe 10% of a 2D desktop screen will almost certainly exceed the 25% spatial area threshold inside an HMD74. Therefore, web-based immersive art requires strict algorithmic throttling to ensure that mathematical visualizations do not exceed the 3Hz strobe limit, guaranteeing the environment remains perceivable, operable, and safe for all users68.

Conclusion

Determining whether to deploy a geometric artwork via a 2D/3D desktop interface or an immersive VR headset requires a rigorous balancing of cognitive benefits, affective impact, performance budgets, and accessibility barriers.

The real value of desktop 3D lies in its accessibility, physiological comfort, and immediate cognitive recall. When the geometric artwork serves as an informational graphic designed for rapid navigation, or when the user requires high text legibility without the ergonomic burden of a heavy headset, the desktop paradigm is superior. It operates within well-understood usability heuristics, minimizes cybersickness to near zero, and allows the user to simultaneously maintain their physical environment context. However, as the topology of the geometric art grows in edge-crossing complexity, the desktop's reliance on monocular depth cues will inevitably result in visual failure.

Conversely, the profound value of immersive headset VR is unlocked when the artwork demands topological disentanglement and aesthetic presence. By providing stereoscopic binocular disparity, depth routing, and head-coupled motion parallax, VR allows the human visual system to effortlessly parse intricate geometric networks comprising thousands of nodes—an order of magnitude greater than what is comprehensible on a desktop. Furthermore, anchored in a local-floor reference space, the artwork transcends the screen to invoke aesthetic awe, spatial presence, and even therapeutic utility in pain distraction.

However, this immersive value is highly conditional. It demands rigorous engineering adherence to 90Hz+ frame rates using WebAssembly to maintain a \<20ms motion-to-photon latency. It requires meticulous spatial typography calculations utilizing WebXR Composition Layers to bypass lens distortion blurring. Finally, it requires strict adherence to XAUR multimodal accessibility inputs and Harding FPA luminance limits to ensure the environment does not induce cybersickness or epileptic distress. Ultimately, depth, scale, and presence add genuine value when the geometric artwork is fundamentally spatial in its conceptual intent—when the structure is too complex to be flattened, and the desired psychological impact requires the user to step out of the role of an observer, and into the role of an inhabitant.

This is for informational purposes only. For medical advice or diagnosis, consult a professional.

Works cited

1. Boost HTML5 Game Performance with WebAssembly \- Playgama Blog, https://playgama.com/blog/2025/04/03/boost-html5-game-performance-with-webassembly/

2. WebXR Browser Support in 2026: What Works, What Breaks, https://www.testmuai.com/learning-hub/webxr-compatible-browsers/

3. What Is the Immersive Web? A Vision for the Next Era of the Internet, https://www.codazen.com/blog/what-is-the-immersive-web/

4. WebVR/WebXR \- Taieb Chaâbini \- Medium, https://taiebchaabini.medium.com/webvr-webxr-ea3941681b4

5. Scalability of Network Visualisation from a Cognitive Load Perspective, https://www.researchgate.net/publication/347759693\_Scalability\_of\_Network\_Visualisation\_from\_a\_Cognitive\_Load\_Perspective

6. INTERACTIVE VISUALIZATION OF LARGE GRAPHS AND, https://graphics.stanford.edu/papers/munzner\_thesis/all.print.pdf

7. A Study of Layout, Rendering, and Interaction Methods for, https://www.computer.org/csdl/journal/tg/2016/07/07390081/13rRUxly8XJ

8. A Study of Layout, Rendering, and Interaction Methods for, https://vis.cs.ucdavis.edu/papers/TVCG\_Kwon2016.pdf

9. VR Reading UIs: Assessing Text Parameters for Reading in VR, https://kaikunze.de/papers/pdf/dingler2018vr.pdf

10. The simulator sickness questionnaire, and the erroneous zero, https://www.frontiersin.org/journals/virtual-reality/articles/10.3389/frvir.2022.945800/full

11. 2D is Better Than 3D \- Nielsen Norman Group, https://www.nngroup.com/articles/2d-is-better-than-3d/

12. Virtual reality design principles: A case study on VRChat \- DiVA portal, https://www.diva-portal.org/smash/get/diva2:1870103/FULLTEXT01.pdf

13. Graphon-Based Visual Abstraction for Large Multi-Layer Networks, https://www.researchgate.net/publication/392839680\_Graphon-Based\_Visual\_Abstraction\_for\_Large\_Multi-Layer\_Networks

14. Information Visualization | Request PDF \- ResearchGate, https://www.researchgate.net/publication/285933806\_Information\_Visualization

15. (PDF) Reevaluating stereo and motion cues for visualizing graphs in, https://www.researchgate.net/publication/221282979\_Reevaluating\_stereo\_and\_motion\_cues\_for\_visualizing\_graphs\_in\_three\_dimensions

16. (PDF) Immersive Collaborative Analysis of Network Connectivity, https://www.researchgate.net/publication/306051854\_Immersive\_Collaborative\_Analysis\_of\_Network\_Connectivity\_CAVE-style\_or\_Head-Mounted\_Display

17. Error rate data from Experiment 1a. Vertical bars represent one, https://www.researchgate.net/figure/Error-rate-data-from-Experiment-1a-Vertical-bars-represent-one-standard-error-about-the\_fig3\_221282979

18. Evaluating Visual Variables in a Virtual Reality Environment, https://www.researchgate.net/publication/367826926\_Evaluating\_Visual\_Variables\_in\_a\_Virtual\_Reality\_Environment

19. Stereoscopic Highlighting: 2D Graph Visualization on Stereo Displays, https://www.researchgate.net/publication/51751980\_Stereoscopic\_Highlighting\_2D\_Graph\_Visualization\_on\_Stereo\_Displays

20. ‪Basak Alper‬ \- ‪Google Scholar‬, https://scholar.google.com/citations?user=kW7fTDQAAAAJ\&hl=en

21. (PDF) Aesthetics and comprehension of curved 3D graphs in Virtual, https://www.researchgate.net/publication/368229755\_Aesthetics\_and\_comprehension\_of\_curved\_3D\_graphs\_in\_Virtual\_Reality

22. Evaluating the Effectiveness of Spatial Memory in 2D and 3D, https://www.researchgate.net/publication/2398144\_Evaluating\_the\_Effectiveness\_of\_Spatial\_Memory\_in\_2D\_and\_3D\_Physical\_and\_Virtual\_Environments

23. Evaluating spatial memory in two and three dimensions, https://www.researchgate.net/publication/222576936\_Evaluating\_spatial\_memory\_in\_two\_and\_three\_dimensions

24. Supporting and Exploiting Spatial Memory in User Interfaces, https://www.emerald.com/fthci/article/6/1/i/1324737/Supporting-and-Exploiting-Spatial-Memory-in-User

25. Spatial Memory of BIM and Virtual Reality: Mental Mapping Study, https://ascelibrary.org/doi/10.1061/JCEMD4.COENG-12808

26. How three-dimensional sketching environments affect spatial thinking, https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0294451

27. Supporting Cognitive Processing with Spatial Information, https://www.cise.ufl.edu/\~eragan/papers/Ragan\_VirtualReality2012.pdf

28. Measuring user preferences and behaviour in a topographic, https://www.tandfonline.com/doi/full/10.1080/17538947.2021.1984595

29. The Effect of Haptic Prediction Accuracy on Presence \- Uni Würzburg, https://downloads.hci.informatik.uni-wuerzburg.de/2018-vr-haptics-presence.pdf

30. Comparison of VR and Desktop Game User Experience in a Puzzle, https://www.researchgate.net/publication/344899877\_Comparison\_of\_VR\_and\_Desktop\_Game\_User\_Experience\_in\_a\_Puzzle\_Game\_Keep\_Talking\_and\_Nobody\_Explodes

31. A survey of challenges and methods for Quality of Experience ... \- PMC, https://pmc.ncbi.nlm.nih.gov/articles/PMC9051501/

32. Sentiment and Storytelling: What Affect User Experience and ... \- MDPI, https://www.mdpi.com/2673-5172/4/1/3

33. Beyond blue-sky thinking: Diurnal patterns and ephemeral, https://www.researchgate.net/publication/366997227\_Beyond\_blue-sky\_thinking\_Diurnal\_patterns\_and\_ephemeral\_meteorological\_phenomena\_impact\_appraisals\_of\_beauty\_awe\_and\_value\_in\_urban\_and\_natural\_landscapes

34. Virtual Reality in Behavioral Neuroscience: New Insights and, https://dokumen.pub/virtual-reality-in-behavioral-neuroscience-new-insights-and-methods-current-topics-in-behavioral-neurosciences-65-303142994x-9783031429941-q-2058850.html

35. XRReferenceSpace \- Web APIs | MDN, https://developer.mozilla.org/en-US/docs/Web/API/XRReferenceSpace

36. A Cross-Cultural Study of Users' Responses to Machine-Generated, https://www.researchgate.net/publication/347546541\_Using\_machine\_learning\_to\_learn\_machines\_A\_cross-cultural\_study\_of\_users'\_responses\_to\_machine-generated\_artworks

37. Hypnotic Enhancement of Virtual Reality Distraction Analgesia, https://pmc.ncbi.nlm.nih.gov/articles/PMC8141382/

38. Hypnotic Enhancement of Virtual Reality Distraction Analgesia, https://pubmed.ncbi.nlm.nih.gov/33724890/

39. HTML in Canvas: Bridging UI and GPU on the Web \- GitNation, https://gitnation.com/contents/html-in-canvas-bridging-ui-and-gpu-on-the-web

40. Building AR/VR with Javascript and HTML \- DEV Community, https://dev.to/orgoldfus/building-arvr-with-javascript-and-html-1ep1

41. Superpowers HTML5 \- The Ultimate Free & Open-Source Game, https://superpowers-html5.com/

42. WebXR Device API \- W3C, https://www.w3.org/TR/2019/WD-webxr-20190205/

43. WebXR Device API \- W3C, https://www.w3.org/TR/2019/WD-webxr-20191010/

44. WebXR Device API \- Spatial Tracking \- Immersive Web at W3C, https://immersive-web.github.io/webxr/spatial-tracking-explainer.html

45. XRSession: requestReferenceSpace() method \- Web APIs | MDN, https://developer.mozilla.org/en-US/docs/Web/API/XRSession/requestReferenceSpace

46. Using bounded reference spaces \- Web APIs | MDN, https://developer.mozilla.org/en-US/docs/Web/API/WebXR\_Device\_API/Bounded\_reference\_spaces

47. A Spatial Web Performance Budget Breakdown — Surface Webb, https://surfacewebb.com/spatial-web-performance-budget-60fps-hidden-costs/

48. Motion-to-Photon Latency Definition \- Unity, https://unity.com/glossary/motion-to-photon-latency

49. User Experience of Reading in Virtual Reality \- Finding Values for, https://arxiv.org/pdf/2004.01545

50. Google Daydream VR Cookbook, https://www.pearson.de/media/muster/ext/9780134845555.pdf

51. (PDF) Legibility of typefaces and preferences of text/background, https://www.researchgate.net/publication/365334957\_Legibility\_of\_typefaces\_and\_preferences\_of\_textbackground\_color\_variations\_in\_virtual\_environment

52. nKast | pixel perf-ect, https://blog.nkast.gr/

53. Tools \- Design Directory, https://www.designdirectory.io/tools

54. 1.2 – Future Text Publishing, https://futuretextpublishing.com/2022/04/12/1-2/

55. WebXR Layers API Level 1 \- W3C, https://www.w3.org/TR/webxrlayers-1/

56. Layers \- xr \- GitHub Pages, https://pmndrs.github.io/xr/docs/tutorials/layers

57. Evaluation of Virtual Reality Games: Simulator Sickness and Human, https://ceur-ws.org/Vol-2246/GHItaly18\_paper\_01.pdf

58. A brief explanation of the Simulator Sickness Questionnaire (SSQ), https://conservancy.umn.edu/bitstreams/70c8fe0f-c84a-4d0e-80ba-b97822a5cf95/download

59. XR Accessibility User Requirements \- XR Design Handbook, https://xrdesignhandbook.com/docs/Others/XR%20Accessibility%20User%20Requirements.html

60. Predicting Motion Sickness in VR \- Fenix, https://fenix.tecnico.ulisboa.pt/downloadFile/1126295043840860/89547-Tiago-Goncalves-dissertacao.pdf

61. Exploring Simulation Sickness in Virtual Reality Pedestrian Scenarios, https://www.mdpi.com/2313-576X/11/3/63

62. The simulator sickness questionnaire, and the erroneous zero, https://research.tilburguniversity.edu/en/publications/the-simulator-sickness-questionnaire-and-the-erroneous-zero-basel

63. Arguing in Favor of Revising the Simulator Sickness Questionnaire, https://pmc.ncbi.nlm.nih.gov/articles/PMC8604025/

64. Cybersickness in Virtual Reality Questionnaire (CSQ-VR) \- TechRxiv, https://www.techrxiv.org/doi/10.36227/techrxiv.21973640

65. Psychometric evaluation of Simulator Sickness Questionnaire and, https://www.semanticscholar.org/paper/Psychometric-evaluation-of-Simulator-Sickness-and-a-Sevin%C3%A7-Berkman/96421c506cd1027552d8040b858948bc351bc973

66. Predicting Motion Sickness in VR, https://fenix.tecnico.ulisboa.pt/downloadFile/1126295043839622/89547-Tiago-Goncalves-resumo.pdf

67. Determining the severity and prevalence of cybersickness in virtual, https://pmc.ncbi.nlm.nih.gov/articles/PMC12139111/

68. WCAG 2.2, Section 508 & EN 301 549 Compliance Matrix Portal, https://epicpaths.com/compliance-matrix/

69. XR Accessibility User Requirements \- W3C, https://www.w3.org/TR/xaur/

70. Introduction to XR Accessibility \- TetraLogical, https://tetralogical.com/blog/2024/09/11/introduction-to-xr-accessibility/

71. The Art of the Possible \- W3C, https://www.w3.org/2021/07/chinese-ig-xr/slides/XR-accessibility.pdf

72. Xaur draft \- Accessible Platform Architectures Working Group \- W3C, https://www.w3.org/WAI/APA/wiki/Xaur\_draft

73. International Guidelines for Photosensitive Epilepsy: Gap Analysis, https://pmc.ncbi.nlm.nih.gov/articles/PMC11872230/

74. Harding test \- Grokipedia, https://grokipedia.com/page/Harding\_test

75. WCAG 2.1 Testing & Remediation Documentation | Custody & Agency, https://custodydesign.com/wcag-testing-reference/

76. 達成基準 2.3.2: 3 回の閃光を理解する, https://waic.jp/translations/WCAG21/Understanding/three-flashes.html

77. Using a tool to ensure that content does not violate the general flash, https://waic.jp/translations/WCAG22/Techniques/general/G15