AI: Summary
This meeting focused on extensive discussions about text presentation in XR/VR environments. The conversation explored fundamental questions about how text should be displayed and interacted with in spatial computing contexts, including debates about rectangles versus non-rectangular containers, the challenges of maintaining readability and context in 3D spaces, and the balance between 2Dlegacy formats and new 3D possibilities. Peter Wasilko demonstrated how he uses Tinderbox as a customizable structure editor for complex file generation. The group discussed practical constraints like the importance of scoped regions for focus, the difficulties of designing for AR where you cannot control the user’s environment, and various technical demonstrations of WebXR implementations. The meeting also touched on the Sloan Foundation-funded work and the development of the Mana menu.
Frode Hegland, Dene Grigar, Mark Anderson, Peter Wasilko, Brandel Zachernuk, Jamie Blustein, Jonathan Finn, Fabien Benetou, Rob Swigart, Ken Perlin, Karl Arthur Smink, Peter Dimitrios, Jimmy Sixdof
AI: Speaker Summary
Frode Hegland served as the meeting facilitator and actively participated in discussions about XR text presentation. He expressed enthusiasm about recent three-person VR sessions with Brandel and Fabian, noting that having three people creates a different, less intense dynamic than two-person interactions. He emphasized wanting to move past debates about PDF versus HTML and rectangles versus non-rectangles, seeking to focus on practical solutions. Frode showed interest in the Mana menu development and requested that participants write brief articles about their vision for headset interactions, starting with “When I put on the headset, I want to be able to reach out and touch XXX and I want to be able to do YYY with it.” He was particularly engaged in discussions about text readability and the practical constraints of current technology.
Dene Grigar co-facilitated the meeting and introduced Mark Anderson as an expert in the hypertext community. She shared personal information about getting a new Apple Watch for running and exercise, using it as a Fitbit equivalent. Dene confirmed Mark‘s excellent paper at hypertext about the augmented self and mentioned her collection of historical software including Web Squirrel, HyperCard, Hypergate, and early calendar software donated by Sarah Smith. She participated in welcoming participants and maintaining the social atmosphere of the meeting.
Mark Anderson provided the main introduction and presentation about his background and Tinderbox. He explained Tinderbox‘s origins and how it evolved from Storyspace, an early hypertext application used before the web. He described Tinderbox as a tool that tries to avoid being totally linear, recognizing that linearity can be an impediment to understanding complex subjects. Mark discussed his paper on the information city concept from the 1990s, which used architectural space as an information exploration mechanism, similar to the Memory Palace idea. He noted his use of multiple screens (two 24-inch monitors and one 14-inch) for productive cross-connections, finding this more effective than single-screen setups. Mark engaged thoughtfully with discussions about VR text presentation and the challenges of maintaining context and readability.
Peter Wasilko provided detailed technical insights about his use of Tinderbox as a customized structure editor. He explained how he uses it to generate complex file formats like LaTeX and Peg parser grammar specifications, offloading the mental work of formatting by creating high-level abstractions with attributes. Peter described his system of templates and agents that handle the complexity of different syntax forms, effectively creating a Lisp-like macro system on top of frozen grammars. He demonstrated deep technical knowledge of parser generators and grammar specifications. Peter also participated in discussions about MIT Media Lab work and historical computing interfaces, mentioning David Small‘s Illuminated Manuscript project and showing interest in how these concepts relate to current XR work. He made a programming joke about needing “Come From’s to get back,” referencing the infamous COMEFROM statement.
Brandel Zachernuk contributed important conceptual insights about the nature of boundaries in XR spaces. He emphasized that the key issue isn’t rectangles versus non-rectangles, but rather how hard the boundary is between regions. He pointed out that printing something outside a page is near-impossible in print but trivial in a VR context, highlighting fundamental differences between traditional and spatial media. Brandel shared a CodePen demo of real models in Vision Pro and engaged with technical discussions about spatial design constraints. He also provided a link to David Small‘s current website, helping connect historical and contemporary work in the field.
Jamie Blustein participated from a position of being overwhelmed with marking (grading) responsibilities. She mentioned trying the app briefly but being too busy to engage fully. Jamie contributed information about Scientific American‘s website changes when bought by a new publisher, noting that they removed hypertext features and sanitized previously published articles. She referenced a demo at HT (Hypertext conference) in Maryland about physical artifacts appearing as 3D in VR, with the interesting capability of modifiers for existing content, using spice or sale examples that added textual annotations to images of airplanes. Jamie also mentioned Molly Bang‘s book about picture composition.
Jonathan Finn participated in discussions about museum interactivity, specifically mentioning the interactive light tables in the Churchill War Museum in London designed by David Small. He found them genuinely cool and effective unlike many museum interactions. Jonathan made an interesting observation that the main room in the Churchill Museum is a single dark room with many things dotted around it, which actually feels like a VR space, showing how physical space design can mirror virtual environment concepts.
Fabien Benetou demonstrated several technical implementations and shared multiple WebXR links during the meeting. He discussed working with default primitives like a three-sided cube with wireframe and transparency as a container for content. Fabien identified intelligibility as his biggest challenge in XR development. He shared demos including a volume primitive container where the container and text within can be moved independently, and a WebDAV browser for filesystems in XR. He also demonstrated console functionality in his XR environments and discussed file format approaches using line-JSON-line-JSON in .stxt files.
Rob Swigart made brief contributions to the conversation, including appreciation for Ken‘s rectangle pun and engagement with the discussion topics.
Ken Perlin contributed wit to the conversation with his comment that “Rectangles are popular in most circles,” which was well-received by the group. He also responded warmly to Mark‘s comment about his mother working with interactive museum technology.
Karl Arthur Smink provided important insights about the inherent challenges of AR development. He pointed out that as a developer or designer, you don’t know half or more of what your user will be able to see, touch, or interact with in their environment. He emphasized how very hard it is to set aesthetic, tone, and mood when you can’t plan around those environmental factors, a crucial constraint that AR developers must work within.
Peter Dimitrios contributed thoughtful observations about scoped regions and focus mechanisms in spatial environments. He discussed how web pages aren’t like pages of a book but there are humane constraints on what we can see in one glance. He noted the need for scoped regions to focus attention, suggesting that zooming into one section of a large diagram or document could work, but acknowledging that we appreciate the 1D storyline that linear book structure provides. Peter mentioned that rectangles are well-known and accepted forms like the golden ratio, 8.5×11, and A4 sizes. He also observed that laptops are strangely shaped like books but we don’t normally hold or use them that way. He asked about Vega-lite in the context of data visualization discussions.
Jimmy Sixdof made brief but relevant contributions, noting that rubrics are big in AI training data now, which relates to how we structure and understand information.
AI: Topics Discussed
Introduction and background of Mark Anderson and his work with Tinderbox
Tinderbox as a hypertext and note-taking tool with non-linear information organization
Storyspace as an early hypertext application and its merger with Tinderbox
The information city concept from the 1990s as architectural information exploration space
Memory Palace as a related concept to spatial information organization
Peter Wasilko‘s use of Tinderbox as a structure editor for generating complex file formats
LaTeX and Peg parser grammar generation through Tinderbox templates
Text presentation and readability in XR/VR environments
The rectangle versus non-rectangle debate for content containers in spatial computing
Boundaries between regions in VR versus traditional print media
AR development challenges when you cannot control the user’s environment
Scoped regions and focus mechanisms for attention management in spatial interfaces
The tension between 2D legacy formats and new 3D spatial possibilities
Multi-screen productivity versus single-screen setups
WebXR implementations and technical demonstrations
Mana menu development (sometimes mistranscribed as Manna menu)
The Sloan Foundation funded project on the future of text in XR
Three-person VR sessions and their different dynamics compared to two-person interactions
Vision Pro capabilities and demonstrations
Physical artifacts appearing as 3D in VR with modifier capabilities
Museum interactivity design, particularly Churchill War Museum examples
Scientific American website changes and removal of hypertext features
Container primitives in XR with wireframe and transparency
File system browsers in XR environments
Data visualization approaches in spatial contexts
Console functionality in XR development environments
File format considerations for XR content (.stxt files with line-JSON structure)
The importance of intelligibility in XR interface design
Golden ratio, 8.5×11, and A4 as established rectangular formats
Linear storyline appreciation despite non-linear possibilities
The relationship between laptop form factors and book-like shapes
AI training data and rubrics as structural frameworks
Interesting Anecdotes
Frode shared a humorous exercise rule about running 15 minutes every day, and if he misses a day, he has to make up the time the next day with an additional 15 minutes, joking that tomorrow he’s scheduled to run for three months.
Dene called out Frode for stealing the running joke from threads, noting it wasn’t his original material, but Frodeadmitted he managed to say it correctly for the first time.
Mark Anderson mentioned that his mother worked with interactive museum technology similar to what David Smallcreated for the Churchill War Museum, providing a personal connection to the historical development of interactive displays.
Peter Wasilko made a programming humor reference about needing “Come From’s to get back,” referencing the notorious COMEFROM statement from programming language history.
Ken Perlin delivered a well-received pun: “Rectangles are popular in most circles,” which got multiple positive reactions from the group.
AI: Concepts Introduced
The information city concept was discussed by Mark Anderson – an idea from the 1990s about using architectural space as an information exploration space, where you would wander around a town and serendipitously discover different areas (like language or science quarters), with buildings representing notes in hypertext that might reveal associations through their spatial juxtaposition.
Scoped regions was a concept emphasized by Peter Dimitrios – the need for defined areas to focus attention in spatial environments, acknowledging that while we can zoom into sections of large diagrams or documents, we still appreciate the 1D storyline structure that linear formats provide.
Hard boundaries versus soft boundaries was a distinction made by Brandel Zachernuk – the key issue in spatial design isn’t whether containers are rectangular or not, but rather how hard or permeable the boundary is between regions, noting that in VR it’s trivial to print something outside a defined page boundary, unlike in physical print media.
Tinderbox as a customizable structure editor was explained by Peter Wasilko – using it to create high-level abstractions with attributes and templates that generate complex file formats, effectively creating a Lisp-like macro system on top of frozen grammars, allowing offloading of mental labor for formatting specifications.
AI: People Mentioned
Mark Anderson (mentioned by Dene Grigar, Mark Anderson himself, and Jamie Blustein as having excellent paper at hypertext conference about augmented self)
Randall (mentioned by Frode Hegland as participating in three-person VR session)
Fabian (mentioned by Frode Hegland as participating in three-person VR session, also present as Fabien Benetou)
Andy Van Dam (mentioned by Mark Anderson as roughly contemporary with Mark at Swarthmore)
Ted Nelson (mentioned by Mark Anderson as from similar catchment as Mark at Swarthmore)
Sarah Smith (mentioned by Dene Grigar as donating early calendar software to her collection)
David Small (mentioned by Peter Wasilko and Brandel Zachernuk for Illuminated Manuscript project at MIT Media Lab, and by Jonathan Finn for Churchill War Museum interactive light tables)
Molly Bang (mentioned by Jamie Blustein in reference to her book about picture composition)
Mark‘s mother (mentioned by Mark Anderson as having worked with interactive museum technology)
AI: Product or Company Names Mentioned
Tinderbox (discussed extensively by Mark Anderson and Peter Wasilko as hypertext and structure editing tool)
Storyspace (mentioned by Mark Anderson as early hypertext app that merged with Tinderbox)
HyperCard (mentioned by Dene Grigar as in her software collection)
Hypergate (mentioned by Dene Grigar as in her software collection)
Web Squirrel (mentioned by Dene Grigar as in her software collection)
Apple Watch (mentioned by Dene Grigar as getting new one for running and exercise)
Fitbit (mentioned by Dene Grigar as comparison for how she uses Apple Watch)
Zoom (mentioned by Dene Grigar as the “magic” behind participants appearing)
Vision Pro (mentioned by Brandel Zachernuk in context of demo with real models)
LaTeX (mentioned by Peter Wasilko as file format he generates through Tinderbox)
Peg parser generator (mentioned by Peter Wasilko as system he creates grammars for using Tinderbox)
Lisp (mentioned by Peter Wasilko as comparison for macro system he created)
MIT Media Lab (mentioned by Peter Wasilko as location of David Small‘s Illuminated Manuscript project)
New Scientist (mentioned by Jamie Blustein in context of not being able to read properly on screen)
Scientific American (mentioned by Jamie Blustein as having removed hypertext features when bought by new publisher)
WebXR (implicit in Fabien‘s demonstrations and links)
WebDAV (mentioned by Fabien in context of browser for filesystems in XR)
CodePen (platform for Brandel‘s demo shared via link)
Vega-lite (mentioned by Peter Dimitrios as potential data visualization tool)
Churchill War Museum / Churchill War Rooms (mentioned by Jonathan Finn as location of David Small‘s interactive light tables)
Swarthmore (mentioned by Mark Anderson as Mark‘s alma mater)
Sloan Foundation (mentioned as funding the project)
AI: Agreements & Disagreements
There was general agreement that the rectangle versus non-rectangle debate needs to move forward, with Frode explicitly stating “NO more PDF V.S. HTML and NO more rectangle V.S. non,” and Brandel refining this to focus on “how hard the boundary is between regions” rather than the shape itself.
Multiple participants agreed with Peter Dimitrios‘s point about needing scoped regions of some sort to focus attention, with Frode and Mark Anderson reacting positively to this observation.
There was consensus around Karl Arthur Smink‘s observation about AR‘s inherent problem that developers don’t know half or more of what users will see in their environment, making it very hard to set aesthetic, tone, and mood, with Frodeand Mark Anderson agreeing this is a fundamental challenge.
Brandel and Peter Dimitrios had a productive exchange about boundaries, with Brandel emphasizing that printing outside a page is trivial in VR but near-impossible in print, adding nuance to the discussion.
There was implicit agreement about the value of Tinderbox as a powerful tool, with both Mark Anderson and Peter Wasilko presenting complementary perspectives on its capabilities without any disagreement.
The group seemed to share appreciation for David Small‘s work, with Jonathan Finn, Peter Wasilko, and Brandel all contributing positive references to his projects at MIT Media Lab and Churchill War Museum.
There was no disagreement about the importance of intelligibility in XR design, with Fabien identifying it as his biggest challenge and others implicitly supporting this through their design discussions.
AI: Other
The meeting had a warm, collegial atmosphere with social banter at the beginning about Zoom timing, Apple Watchuses, and exercise habits, showing the group’s comfort with each other.
Jamie Blustein was unable to fully participate due to being overwhelmed with marking (grading) responsibilities, which she couldn’t fully describe because the meeting was being recorded – suggesting there may be challenging institutional circumstances.
The transcript includes both spoken conversation and a detailed chat log showing active text-based participation alongside the verbal discussion, with participants sharing links, reactions, and additional context.
Frode made a specific request twice in the chat for participants to write brief articles starting with “When I put on the headset, I want to be able to reach out and touch XXX and I want to be able to do YYY with it,” indicating this is an important next step for the project.
The meeting demonstrates the interdisciplinary nature of the Sloan Foundation project, bringing together people with expertise in hypertext history, software development, XR implementation, museum design, and academic research.
There were several technical demonstrations during the meeting, with Fabien sharing multiple working links to WebXRimplementations that participants could access during and after the meeting.
The discussion touched on historical computing concepts (like COMEFROM statements and Memory Palace) and connected them to contemporary XR development, showing awareness of how past ideas inform current work.
The group’s consideration of rectangles, A4 paper, and golden ratios shows they’re grappling with how established human interface conventions translate (or don’t translate) to new spatial computing contexts.
Chat Log URLs
Chat Log Summary
The chat log shows active parallel discussion alongside the spoken conversation. Participants shared reactions through emojis and brief comments. Frode emphasized wanting to move past PDF versus HTML and rectangle versus non-rectangle debates. Peter Dimitrios contributed important observations about scoped regions and the need for humane constraints on what can be seen at a glance, noting that rectangles are well-known accepted forms (golden ratio, 8.5×11, A4). He observed that laptops are strangely shaped like books but we don’t use them that way. Ken made his well-received pun about rectangles being popular in circles. Brandel emphasized that the issue is less about rectangles versus non-rectangles and more about how hard boundaries are between regions, noting that printing outside a page is trivial in VR but near-impossible in print. Karl Arthur Smink made the crucial point about AR‘s inherent problem of not knowing what the user’s environment contains. Fabien shared multiple technical links to his WebXR implementations, including volume primitives and WebDAV browsers, and discussed working with three-sided cubes with wireframes as containers. Jimmy Sixdof noted that rubrics are big in AI training data. Peter Wasilko made the COMEFROM programming joke. Jamie discussed Scientific American removing hypertext features and mentioned HT Maryland demos and Molly Bang‘s work. Jonathan Finn shared information about Churchill War Museum interactivity. Mark Anderson noted frustrations with multi-screen versus single-screen productivity and made jokes about “no JavaScript JSON.” The chat shows a technically sophisticated group actively sharing resources and building on each other’s ideas in real-time.
Issues to Get Back To
Frode made a request twice for participants to write brief articles or paragraphs starting with “When I put on the headset, I want to be able to reach out and touch XXX and I want to be able to do YYY with it,” suggesting this is a pending action item that participants should follow up on.
The intelligibility challenge that Fabien identified as his biggest problem suggests ongoing work that needs to be addressed in future sessions.
The discussion about AR‘s inherent challenges where developers can’t control the user’s environment (raised by Karl) seems like a fundamental issue that will require continued exploration and potential solutions.
The balance between maintaining linear storyline benefits while exploring non-linear spatial possibilities (discussed by Peter Dimitrios and others) appears to be an ongoing design challenge without a clear resolution yet.
Things Readers Might Miss
The significance of this being Sloan Foundation funded work, which contextualizes the research and development discussions as part of a formal grant project focused on the future of text in XR.
The Mana menu (not Manna menu) is a specific project the group is working on, which wasn’t discussed in detail but is clearly an important ongoing element of their work.
Dene Grigar maintains a historical software collection that includes Web Squirrel, HyperCard, Hypergate, and early calendar software, suggesting she serves as something of an institutional memory for the hypertext community and its tools.
The casual mention of three-person VR sessions with Randall and Fabian suggests the group is actively experimenting with multi-user VR collaboration as part of their research methodology.
Mark Anderson‘s work with Tinderbox has been ongoing since at least 2004 (when Mark Anderson started using it), and the tool itself has roots going back to Storyspace from before the web existed, providing important historical context for hypertext development.
Peter Wasilko‘s deep technical implementation of Tinderbox for parser generation shows the tool’s flexibility extends far beyond its designed use cases, which may inspire others to think creatively about how to use spatial computing tools.
Jamie‘s being overwhelmed with marking and the fact she couldn’t explain everything because the meeting was being recorded hints at institutional pressures on academics that affect their ability to fully participate in research collaborations.
The group’s comfort with making jokes and puns (like Ken‘s rectangle joke) while discussing serious technical matters shows this is a collegial community that values both rigor and humor.
Jonathan Finn‘s observation that the Churchill Museum main room feels like a VR space suggests that physical space design and VR design may inform each other in ways that aren’t immediately obvious.
The fact that Scientific American removed hypertext features when changing publishers (mentioned by Jamie) reflects broader industry tensions about web versus traditional publishing that may affect how text is presented in any medium, including XR.
The discussion about rectangles being tied to well-known formats (golden ratio, 8.5×11, A4) shows the group is aware they’re not just designing for technical constraints but for deeply ingrained human expectations and centuries of print culture.