Home onD'Arcy Norman, PhD

Experimental Obsidian 3D Graph Renderer

nospam-dnorman@me.com (D'Arcy Norman) — Fri, 10 Apr 2026 13:05:56 -0600

I’ve been playing around with some ideas for visualizing the notes and connections in my Obsidian vault, and how to go beyond just producing pretty pictures and into something that might be useful to make sense of the whole thing. It’s not quite there yet, but it’s getting close. Vibecoding this with Claude Code has let me take some ideas I’ve had for years, and actually building a thing in a couple of hours.

Here’s a 30-second timelapse of my notes, sorted by date created. The first notes that show up are related to my MSc and PhD programs, and then it really starts to fill in once I started using Obsidian for everything back in 2022.

A video is embedded here, but may not be displayed on a preview or if your browser does not support the video tag.

I’m working on more tools to filter the nodes (by date, tag, folder), as well as being able to select a single note and visualize its connections (over a number of “hops” to the rest of the graph).

Here’s the graph subset based on the “UCalgary Microcredentials Framework” note, +2 hops:

Screenshot of the Graph Render Obsidian plugin

Rendering setting panel open to the left, with UCalgary Microcredentials Framework note selected and notes up to 2 hops away displayed.

Image by D'Arcy Norman

The plugin uses WebGL and is hardware acceleration enabled. The initial layout takes a couple of seconds, and then it’s a fully interactive 3D graph environment.

From Claude’s Readme notes:

AI-Generated Content via Claude Code (Sonnet 4.6)

3D Graph Render — Obsidian Plugin

A hardware-accelerated 3D visualisation of your Obsidian vault’s note graph. Built on Three.js + WebGL, routed through Metal on macOS for full GPU acceleration. Renders in a Galaxy aesthetic: deep-space amber nebula, glowing clustered nodes, and flowing curved edge filaments.

Features

Rendering — Galaxy Style

3D force-directed layout — d3-force-3d simulation runs inside the animation loop and self-organises automatically; watch the graph settle in real time
Instanced rendering — single GPU draw call for all nodes regardless of vault size
Additive bloom — nodes use additive blending so dense clusters self-brighten into hot white cores; UnrealBloomPass adds soft halos
Procedural nebula background — a full-sphere FBM (fractional Brownian motion) GLSL shader renders animated amber/orange gas clouds behind the graph
Curved Bezier edges — at layout convergence, straight edges are replaced by 8-segment quadratic Bezier curves with a perpendicular bow, giving a galactic-arm filament appearance
Clustered particle haze — 6 000 warm golden particles concentrated near the graph centre via power-law radial distribution
Cluster-coloured nodes — coloured by tag group or folder; all nodes get a colour via deterministic hashing (nothing appears grey)

and

AI-Generated Content via Claude Code (Sonnet 4.6)

Filters (Filter button → right panel)

All filters use a Seed + Expand model: matching notes are seeds (full size/brightness), their direct neighbours are shown at reduced size/brightness, everything else fades to near-invisible.

Filter	What it does
Date	Seeds = notes created or modified within last day / week / month / year
Tags	Seeds = notes containing any of the selected tags (checkbox list, top 50 by frequency)
Folder	Seeds = all notes inside the selected folder
Starting note	Seeds = a specific note + its N-hop neighbourhood (hop depth 1–5)

Filters compose: all active conditions are unioned into one seed set.

A match count at the bottom of the panel shows N matches · M neighbors.

Note on date filters: “Modified” reflects the file system mtime, which is updated by sync tools (iCloud, Obsidian Sync, Dropbox) even without manual edits. Use a tighter range (Last day) for precise results.

Search

Type in the Search box for real-time highlighting:

Matching nodes glow bright yellow and grow 50% larger
Non-matching nodes shrink and dim to near-invisible
Matches on note name, tags, and folder path

digital transformation is a teaching and learning opportunity

nospam-dnorman@me.com (D'Arcy Norman) — Tue, 07 Apr 2026 10:49:58 +0000

Reposting the UCalgary News post I wrote for our TI Newsletter:

Digital Transformation (Dx) in higher education has largely been framed as an organizational and IT challenge: enterprise systems, cloud migration, data governance, workforce development. That framing makes sense for the people leading those efforts, and the work is genuinely important. But it can leave a gap for those of us whose daily work is about teaching and learning. What does digital transformation look like when the starting point is pedagogy, not infrastructure?

EDUCAUSE draws a useful distinction between three things that often get conflated. Digitization is converting analogue to digital (scanning syllabi, recording lectures). Digitalization is using digital tools to change existing processes (moving a course into an LMS, accepting online submissions). Digital transformation is something deeper: coordinated shifts in culture, workforce, and technology that enable fundamentally new educational models.

By that definition, most of what happens in teaching and learning support is digitization or digitalization. We help instructors use tools. We move things online. We make existing processes more efficient. That’s valuable work, but it’s not transformation.

So what does Dx look like when it’s centred on teaching and learning? A few examples stand out.

Arizona State University spent years building faculty trust in adaptive courseware before scaling it across gateway courses. The results were significant: a 24% improvement in pass rates in introductory biology and a 90% reduction in dropouts. But the technology wasn’t the hard part. The hard part was redesigning courses around personalized learning pathways and convincing faculty to let go of the lecture. As their adaptive programme manager put it: “You can’t do this a little bit.”

Georgia State University used predictive analytics not as an IT project but as a lever for institutional transformation. They track 800 risk factors for 50,000 students daily, but the analytics only work because they hired dozens of advisers to act on the alerts. The six-year graduation rate went from 32% to over 54%, and achievement gaps across race and income were eliminated. Their framing is instructive: the goal was not to change the nature of the students but to change the nature of the institution.

In the UK, Jisc co-developed a digital transformation framework with nine sector bodies, including Advance HE and the Association for Learning Technology, that explicitly focuses on people and practices, not just processes and technology. They piloted it with 24 universities in 2024, producing case studies that connect infrastructure changes to pedagogical improvement.

Closer to home, eCampusOntario offers micro-credentials specifically for educators navigating Dx, and BCcampus recently defined 21 digital learning competencies organized into mindsets, skillsets, and toolsets. Both position faculty capacity as the foundation of transformation.

The common thread across all of these? The most successful teaching-and-learning Dx initiatives are not primarily about technology adoption. They involve pedagogical redesign, faculty development, and organizational culture change. Technology is necessary but not sufficient.

For those of us working in teaching and learning centres, this is both validating and challenging. It means the work we do (supporting course design, building faculty capacity, fostering communities of practice) is central to digital transformation, not peripheral to it. But it also means we need to ask ourselves where we’re digitizing when we could be transforming.

At UCalgary, these questions are live. The university has a Digital Transformation and Technology portfolio reporting to the President, and the Ahead of Tomorrow strategic plan includes an Academic Innovation Plan that addresses the teaching and learning dimensions of institutional change. The infrastructure side of Dx is well underway. The question is how deeply the teaching and learning side keeps pace.

The Taylor Institute is hosting the 2026 Conference on Postsecondary Learning and Teaching on exactly this topic: “From Disruption to Connection: Digital Transformation in Postsecondary Education.” The conference framing asks how digital transformation can move beyond disruption to become a force for connection, linking students, educators, communities, and knowledge. It asks what relational approaches to Dx might look like, and what skills, mindsets, and supports educators and students need to navigate technology-rich environments confidently and ethically.

Those are the right questions. And they’re questions that start with pedagogy, not infrastructure. Which is, I think, where the most interesting work is going to happen.

Linkblogging

nospam-dnorman@me.com (D'Arcy Norman) — Sun, 05 Apr 2026 16:12:05 -0600

All of my links for the last few months have been saved in my bookmarks tool, and I’ve been finding that useful. I’ve also appreciated Tom Woodward and Grant Potter and others for sharing their weekly links. My previous attempt - the “Weekly Reflections” thing - kind of petered out, but I just implemented a new Section on my site to post weekly linkblogs. The content is generated using my bookmark tool’s Archive feature, generating a markdown post containing all links in a given week. Paste that into a new Links post here on my blogthing, and hey presto I’m linkblogging. Maybe this will nudge me to better annotate my bookmarks. The added bonus of posting weekly linkblogs - they’ll be included in my handy dandy search tool so I don’t have to remember “did I write about it or did I just bookmark a thing?”

Linkblog posts aren’t included on the front page or main RSS feed because they’d swamp both (similar to Photos).

The weekly linkblog things will also be posted to my Mastodon account (which now has 3 RSS feeds from my website posting new items - the main combined feed, photos, and linkblogs)

I also took the opportunity to build out a Sections thing on my website, and refactored the main navbar because the “Archives” menu was turning into a dumping ground.

Anyway. Will I keep posting them? I think so - with the bookmarks tool’s Archive tool making it super easy to grab markdown of a week’s bookmarks? Can’t get much easier than that…

How I built the Typeset custom theme for this website

nospam-dnorman@me.com (D'Arcy Norman) — Sun, 29 Mar 2026 19:45:00 +0000

I’ve been running this site on Hugo for several years now, and for most of that time I used other people’s themes and tweaked things around the edges. A few months ago, I decided to build my own theme from scratch with Claude Code. The theme is called Typeset.

Why build a custom theme?

The short answer: I got tired of fighting someone else’s decisions. Every borrowed theme comes with assumptions about what a blog is and what it should look like. Eventually those assumptions start getting in the way.

The longer answer: I wanted to actually understand my own site. 24 years of content across eight different content types, and I wanted layouts, CSS, and templates that I understood well enough to fix when something broke. Building it myself - even with Claude doing the heavy lifting on actual code - meant I had to understand every decision. And it gave me the flexibility to try new ideas - like the new “Series” feature I implemented this week (using Hugo’s flexible taxonomies and some partials and layouts).

I’d built a simpler theme almost exactly a year ago - and it mostly worked well, but I could never get it to fully do what I had in mind and the workflow was not great (using the Claude web application to build files that I downloaded and edited into place). So I decided to start over and build a full(er) Hugo theme from scratch using Claude Code.

The Series page

Screenshot of the Series page, in 1bit css theme

Image by D'Arcy Norman

Why ’typeset'?

It started with a goal of building a theme that felt like a properly designed and highly readable website with good typography and layout. Most of the CSS themes (see below) currently use Helvetica because I like the feel of it (and my dissertation was set in Helvetica, despite the objections of my PhD supervisor, so I guess I’m committing to the bit).

CSS themes

The most visible feature (that most people probably won’t see) of Typeset is the theme switcher in the header. There are seven selectable colour themes plus a boid/flocking animation overlay for the terminally whimsical¹:

Light - clean and minimal (auto toggle based on browser display mode)
Dark - dark grey with cool blue accents (auto toggle based on browser display mode)
Terminal - amber text on near-black, Departure Mono font, subtle CRT scanline effect
Mid-Century - warm tan and caramel tones, a nod to 1950s design
1-Bit - pure black and white, ChicagoFLF font, thick borders - basically a Mac SE in 1987
Foothills - sky blue and earthy brown, inspired by a photo of the Alberta landscape I spend a lot of time looking at (but mostly to test my CSS Colour Themer utility’s ability to create colour palettes)
1975 - midnight navy with newsprint white content panels and NASA-worm red accents; OMNI magazine vibes

If you’ve selected a theme, your choice persists across visits via localStorage. If you haven’t picked one, the site respects your OS-level prefers-color-scheme setting and defaults to light or dark accordingly. I’ll probably continue to add and remove themes to try stuff out.

Each theme is a separate CSS file with all colours and fonts scoped to html[data-theme="..."] attributes. The main stylesheet defines a shared token system - spacing, type scale, container widths - that all themes inherit. The theme-specific files only override what they need to.

One thing that took more work than expected: preventing a flash of the wrong colours on page load (FOUC). The site uses an inline script in the document <head> that reads localStorage and sets the data-theme attribute synchronously, before the browser paints anything. There’s also a small block of critical CSS inlined in the base template as a fallback. It mostly works.

Content types

This site has always had more than just “blog posts.” Typeset has dedicated templates for eight content types:

Posts - long-form writing, what you’re reading now
Notes - short microblog-style observations (mostly used during my MSc and PhD programs)
Asides - link posts and quick reactions
Photos - images with captions, presented in a CSS grid
Reflections - weekly journals: work, links, watching, reading
Podcast - episode pages with embedded audio
Pages - static content (about, blogroll, etc.)
Consulting - professional services, deliberately minimal

Each type has its own archetype template so hugo new notes/my-note.md gives you the right front matter pre-filled. The section pages use different list layouts - photos get a grid, notes and asides get year-grouped lists, posts get paginated cards.

Obsidian as CMS application

My writing workflow lives in Obsidian. I open my website’s content directory as a Vault in Obsidian, and just use built-in Obsidian tools (templates, plugins, etc.) to manage content. I draft posts there, using note templates I’ve built up over time. The content/templates/ directory holds those Obsidian templates; it’s mounted but explicitly excluded from the Hugo build so the templates never end up as published pages. It’s a small thing, but it means my publishing tool and my writing tool share the same file tree without stepping on each other.

The typeset theme files

Screenshot of a macOS Finder window showing the theme file structure.

Image by D'Arcy Norman

Server-side integrations

Hugo generates static HTML, but this site has a few pieces that require a real server:

Search - Earlier in the site’s history, search used a JavaScript solution that downloaded a JSON index of all content and ran queries in your browser. That was fine until the index grew to 7–8MB and started taking several seconds to load before you could search anything. The current search is a SQLite database built at deploy time from Hugo’s JSON output, with a PHP API endpoint serving queries. I wrote about how that was built last year. (Source code available on GitHub)

Comments - A custom PHP comment system with SQLite storage. (Source code available on GitHub)

Bookmarks - A PHP bookmark management system, also server-side. (Source code available on GitHub)

Mastodon - After each deploy, a shell script posts new content to my Mastodon account at @dnorman@cosocial.ca. (Source code available on GitHub)

How it was built

Claude Code did most of the actual coding. I drove the design decisions - what content types to support, how indices should work, what the theme switcher should feel like, which of the eight themes to include, how I wanted menus and layouts to work - and Claude translated those into working Hugo templates and CSS. The process was iterative and often involved me explaining what was wrong with the output and asking for adjustments.

Is it “vibecoding”? Sort of. I had to understand enough Hugo’s templating system and CSS custom properties to have useful conversations about what I wanted. Claude can’t read my design ideas out of thin air. But I wouldn’t have shipped this without it - the amount of template code involved would have taken months on my own.

I hadn’t planned to share the source code, given how custom this theme is. But maybe parts of it could be handy? Anyway: Source code is available on GitHub.

one of the themes has something of an easter egg that adds to the boids/flocking/connections background… ↩︎

DICE in Practice: Adopting Generative AI Guidelines

nospam-dnorman@me.com (D'Arcy Norman) — Fri, 27 Mar 2026 15:00:00 -0600

The scenario

A mid-sized research university is developing its institutional response to generative AI. There is no single “AI project” - instead, there are dozens of overlapping conversations happening at different levels: individual instructors experimenting with ChatGPT in their courses, departments writing local policies, a provost-level working group drafting institutional guidelines, and national disciplinary organizations publishing position statements. Some faculty are enthusiastic, some are anxious, and most are somewhere in between.

This is a classic multi-level change initiative, and it’s exactly the kind of situation where the DICE Framework is useful - not because it simplifies the complexity, but because it helps people locate themselves within it.

Mapping the scenario

The table below maps the four DICE modes (Decide, Influence, Contribute, Engage) across the four organizational levels (Micro, Meso, Macro, Mega) for this scenario. Each cell describes how a person might participate at that intersection. No single person occupies all sixteen cells - the point is to identify where you are and whether your energy is well-placed.

	Decide	Influence	Contribute	Engage
Mega	Vote on national disciplinary standards for AI use (e.g., accreditation body)	Shape sector conversation through professional networks, conferences, publications	Present institutional case studies at conferences; contribute to cross-institutional research projects	Attend sector-wide events on AI in higher ed; follow emerging policy and literature
Macro	Approve institutional AI guidelines in a governance body (e.g., academic council)	Serve on the provost’s AI working group; advocate for specific policy directions	Pilot AI tools in courses and share results with the working group; write resource guides for colleagues	Attend town halls on AI policy; read institutional communications; complete AI literacy training
Meso	Set department-level expectations for AI disclosure in course outlines	Facilitate faculty discussions on AI and assessment; build informal coalitions across programs	Run a reading group on AI and pedagogy; document and share what’s working in your program	Participate in a faculty learning community on AI; attend a colleague’s workshop
Micro	Decide whether and how to permit AI use in your own courses; set expectations in your syllabus	Have one-on-one conversations with colleagues about their AI approaches	Experiment with AI-assisted assignment design; document your process and share with your department	Try an AI tool for the first time; attend an introductory workshop; read a blog post about AI and assessment

Reading the matrix

A few things to notice:

Most people are in the bottom-right quadrant. If you’re an instructor who attended a workshop on AI and is now trying ChatGPT in your course design process, you’re engaging at the mega and macro levels and contributing at the micro level. That’s entirely appropriate - and it’s meaningful participation in this change initiative, even though you’re nowhere near the governance table.

The diagonal is a common pattern. Many people decide at micro, influence at meso, contribute at macro, and engage at mega. This isn’t a rule - it’s just a frequent shape that emerges because formal authority tends to be strongest close to your own practice and weakest at the broadest levels.

Meso is where initiatives stall or scale. The department or faculty level is where individual experiments either get picked up by colleagues or stay isolated. An instructor who has redesigned their assessments for an AI context (contributing at micro) can only scale that work if there’s a meso-level mechanism - a learning community, a departmental conversation, a curriculum committee - that picks it up. Without meso, macro-level policy and micro-level innovation never connect.

Frustration points are predictable. A faculty member on the provost’s AI working group who tries to decide institutional policy will be frustrated - they can influence at that level, but the decision authority belongs to the governance body. An associate dean who spends all their time engaging (reading about AI, attending webinars) when their role gives them the authority to decide department-level expectations is underusing their position. The framework makes these mismatches visible.

Using this for your own context

You can build a DICE matrix for any change initiative. Pick your scenario - a new assessment policy, a platform migration, a curriculum redesign, an accessibility initiative - and ask:

Who can Decide at each level? These are the people with formal authority. There are usually fewer of them than you think.
Who can Influence? These are the committee members, the facilitators, the coalition-builders. They shape decisions without making them.
Who can Contribute? These are the doers - the people experimenting, documenting, creating, modelling. Their work generates the evidence that informs decisions.
Who is Engaging? These are the people learning, attending, reading, building understanding. They’re the next wave of contributors, and their receptive participation is what makes a community of practice function.
Where are the gaps? If nobody is contributing at the meso level, your initiative has a scaling problem. If nobody is engaging at the macro level, your institutional policy may lack broad awareness. If someone is trying to decide at a level where they can only influence, they’re headed for burnout.

The goal isn’t to fill all sixteen cells, but rather to see the pattern clearly enough to spend your energy where it matters.

Introducing the DICE Framework for Higher Ed Change Leadership

nospam-dnorman@me.com (D'Arcy Norman) — Mon, 23 Mar 2026 16:15:00 +0000

Leading change in a university involves many kinds of participation at once. In any given week, the same person might chair a committee with real decision-making authority, serve in an advisory role on another, pilot something new in their own teaching, and sit in a conference session absorbing ideas they hadn’t considered before. Each of those is a different kind of engagement, and each contributes to change in a different way.

Over the past several months, working through ideas with colleagues and drawing on research from multiple fields, I have been developing a framework that I am calling DICE: Decide, Influence, Contribute, Engage. This post describes its structure, theoretical foundations, and potential applications. It is a way of mapping how people participate in and lead educational change - not as a hierarchy, but as four distinct modes of engagement that operate across different organizational levels.

The Problem

Several established frameworks describe how individuals relate to change and to systems of authority, but each addresses only a portion of the problem.

Covey (1989) proposed a model of three concentric zones - the Circle of Control, the Circle of Influence, and the Circle of Concern - to help individuals focus their energy where they have the most agency. This model has been widely adopted in both personal development and organizational contexts, and for good reason: the core insight, that energy spent worrying about things outside one’s influence is largely wasted, is a useful one. However, Covey’s model treats the outermost zone - “concern” - as essentially passive. A person attending a conference, learning from colleagues, or staying informed about sector-wide developments is doing something more than worrying. They are participating in a community of practice, and this participation has value that Covey’s framing does not capture.

Arnstein (1969) offered a different model: a ladder of citizen participation, ranging from manipulation at the bottom to citizen control at the top. Arnstein’s ladder has been influential in fields ranging from urban planning to public health, but it arranges forms of participation into a strict hierarchy in which “higher” is always better. This does not describe what happens in universities, where a person may appropriately hold formal authority in one context and participate as a learner in another - and where both of those forms of engagement are doing real and necessary work.

Paulhus (1983) developed the concept of “spheres of control,” distinguishing between personal, interpersonal, and sociopolitical domains of perceived control. This work usefully expanded Covey’s single dimension into multiple domains, but remained focused on individual psychology rather than organizational positioning. Paulhus and Van Selst (1990) further validated this multidimensional structure, but the framework does not address the organizational levels at which change is enacted.

What is missing across these models is a framework that treats multiple forms of participation as legitimate and valuable, and that maps them against the specific organizational levels where educational change actually happens. DICE is an attempt to address this gap.

The Four Modes

The DICE Framework identifies four modes of leadership and participation. These are not stages in a hierarchy but distinct ways of relating to a change initiative, each appropriate in different contexts.

Decide. The Decide mode describes situations in which a person holds formal authority to make decisions that are implemented. This is the committee chair, the department head, the person who signs off on a platform contract. Decide draws on the organizational control literature (Eisenhardt, 1985; Ouchi, 1979) and on Paulhus’s (1983) concept of spheres of control, adapted here to describe organizational positioning rather than individual psychology. Formal decision-making authority matters - but it is not the only form of participation that matters.

Influence. The Influence mode describes situations in which a person can shape outcomes through persuasion, facilitation, and coalition-building, but does not hold final authority. Research on distributed leadership in higher education (Harris, 2008, 2022; Kezar & Holcombe, 2017) describes how influence operates through networks and relationships rather than through positional authority. The advisory committee member who frames the question that reorients a discussion is exercising influence, even when the formal decision lies elsewhere.

Contribute. The Contribute mode describes situations in which a person adds value through doing: modelling practices, documenting what works, creating resources, running pilots, gathering evidence. Fairman and Mackenzie (2012) describe this as “contributory leadership” - expanding spheres of teacher leadership action from individual practice outward to broader community engagement. York-Barr and Duke (2004) similarly document how teachers lead through modelling and demonstrating rather than through formal authority. In my experience, the Contribute mode is where the most visible innovation occurs in universities, though it often goes unrecognized as leadership.

Engage. The Engage mode describes situations in which a person participates to learn, to be informed, and to be influenced. This mode is the easiest to dismiss as passive, but it is not. Wenger’s (1998) concept of legitimate peripheral participation, developed initially by Lave and Wenger (1991), reminds us that learning within a community is itself a form of meaningful participation. Workshop attendance, conference learning, reading the field - these are active choices that build the capacity for everything else. The Engage mode is what distinguishes DICE from an earlier three-mode model I was developing; its addition was motivated by the recognition that Covey’s “concern” zone and its derivatives consistently undervalue receptive participation.

The Organizational Levels

DICE operates across four organizational levels drawn from the 4M Framework developed by Simmons (2016) within the Scholarship of Teaching and Learning community. The 4M model - itself influenced by Bronfenbrenner’s (1979) ecological systems theory - identifies four nested levels at which educational change takes place.

Micro: the individual. One’s own practice, one’s direct reports, one’s immediate work. This is where most people spend most of their time, and where they typically have the most agency.

Meso: the department, the faculty, the learning community, the working group. Roxå and Mårtensson (2009) documented how informal “significant networks” at this level shape teaching practice in ways that formal structures often do not. The meso level is the critical connector: it is where individual innovations are amplified, where institutional policies are translated into practice, and where networks sustain change over time. In my experience, meso is the most important level in universities, and the most consistently underinvested.

Macro: the institution. Policies, platforms, governance structures, strategic plans. This is the level at which decisions affect everyone, and at which the gap between decision-making and implementation is often widest.

Mega: beyond the institution. Disciplinary communities, national conversations, sector-wide initiatives. Most people engage at this level rather than decide at it, and that is entirely appropriate.

The 4M Framework has been further developed through empirical work examining how SoTL practice operates across these levels (Simmons, 2020) and through institutional applications demonstrating that teaching and learning initiatives require coordinated action at multiple levels simultaneously (Kenny, Eaton, Watson, & El-Hage, 2021).

The 4×4 Matrix

The contribution of the DICE Framework lies in the integration of these two dimensions. Combining four modes of participation with four organizational levels produces a 4×4 matrix - sixteen positions in which a person might find themselves participating in educational change. The power of the framework is not in any single cell, but in the pattern.

	Decide	Influence	Contribute	Engage
Mega	Set sector-wide standards (rare)	Shape field through professional networks	Publish, present, contribute to sector knowledge	Attend conferences, read field literature
Macro	Vote on institutional policy in governance bodies	Advise through committees and working groups	Gather evidence, run pilots, provide feedback	Stay informed on institutional strategy
Meso	Set department or program direction	Facilitate communities, build cross-unit coalitions	Share innovations, model practices, document what works	Participate in faculty discussions and learning communities
Micro	Determine your own priorities and practices	Shape colleagues’ thinking through conversation	Experiment, create resources, innovate in your work	Attend workshops, learn new tools and approaches

To make this concrete: in my own role leading a learning technologies team at a teaching centre, I simultaneously Decide at the micro level about my team’s priorities and practices, Influence at the meso level through cross-unit collaboration and advisory work, Contribute at the macro level by gathering evidence and running pilots, and Engage at the mega level by attending conferences and reading field literature. That is four different modes operating at four different levels at the same time. The framework makes this visible - and more importantly, it helps identify situations in which energy is being spent trying to Decide at a level where one can only Influence, or trying to Influence where the most productive mode would be to Contribute or Engage.

An instructor’s pattern would differ: they might Decide at the micro level about the design of their course, Influence at the meso level through conversations with departmental colleagues, Contribute at the macro level by sitting on an institutional committee, and Engage at the mega level by attending conferences and reading field literature. The matrix gives both people - the team leader and the instructor - a shared language for describing what they do, without implying that one pattern is more valuable than the other.

Developmental Pathways

The DICE Framework describes a natural developmental progression: Engage → Contribute → Influence → Decide. New faculty members or staff typically begin by engaging - attending workshops, learning institutional culture, building understanding. Over time, they begin contributing through their own practice. With experience and network-building, they move into influence. Some eventually take on formal decision-making roles. This trajectory aligns with research on developmental pathways in teaching and learning leadership (Moore, Felten, & Strickland, 2018) and with Wenger’s (1998) description of movement from peripheral to full participation in communities of practice.

This progression is not mandatory, nor is it always linear. A person may Decide at the micro level while Engaging at the mega level - and this is not a failure but an appropriate matching of mode to context. The framework explicitly resists treating Decide as the goal and all other modes as lesser stages on the way there.

The Relationship to RACI

The RACI matrix (Responsible, Accountable, Consulted, Informed) and its variant RASCI (adding Supportive) are widely used in project management to assign roles for specific tasks. The framework has no single scholarly origin; it emerged from corporate practice in the 1970s and 1980s, variously attributed to work at DuPont Corporation and Ernst & Young. Despite its lack of formal academic grounding, RACI has become a standard tool in institutional governance, including in universities, and I have used it many times.

RACI and DICE are doing fundamentally different things, and understanding the distinction matters.

RACI is task-scoped. It answers the question: for this specific deliverable or decision, who plays which role? It is typically applied at a single organizational level - a project team, a committee, a working group - and maps people to discrete responsibilities. When one builds a RACI for a learning management system renewal process, one is clarifying who signs the contract, who writes the RFP, who is consulted, and who is informed.

DICE is positional and multi-level. It answers a different question: given where a person sits in an organization, what mode of participation is available to them at each level, and are they spending their energy accordingly? DICE is not about who does what on a task - it is about understanding the nature of one’s agency across the full landscape of an initiative or a role.

A concrete example may clarify the difference. A RASCI matrix for course experience survey governance might show that the Associate Director of Learning Technologies (yours truly) is Supportive on survey design and Consulted on reporting access. This is helpful for task coordination. But it does not describe how that Associate Director should think about their relationship to the initiative across organizational levels. DICE does: at the macro level, they might Influence survey governance through advisory input; at the meso level, they might Contribute by building faculty capacity to interpret survey data; at the micro level, they Decide how their own team administers and supports the survey tool.

The two frameworks are complementary, not competing. RACI helps organize a well-structured project. DICE helps one understand where one stands in relation to an initiative - and whether one is expending energy in the right mode at the right level. In my experience, most frustration in university committees comes not from unclear RACI assignments, but from people operating in the wrong DICE mode: trying to Decide when they can only Influence, or trying to Influence when the most productive thing would be to Contribute or Engage.

Using DICE for Planning

The framework is most useful as a diagnostic and planning tool. For any initiative, one can ask:

What organizational level am I working at?
What mode can I realistically operate in at this level?
Is this the right mode for what I am trying to accomplish?
Where am I spending energy trying to Decide when I can only Influence - or trying to Influence when I should be Engaging?
Who on my team or in my network can operate in complementary modes?
Are there gaps? Where are they, and who can help to fill them?

For teams, the framework helps map coverage: who can Decide, Influence, Contribute, and Engage at each level? Where are the gaps? Where is the team over-indexed? Is anyone burning out trying to operate in a mode that does not match their actual authority?

Future Directions

DICE is a work in progress. I am developing it through a combination of theoretical grounding and practical application in my own institutional context. I have begun using it as a planning tool with my team and as a lens for understanding how digital learning initiatives move - or stall - across organizational levels.

Future work will explore specific applications: how DICE maps to platform governance, how it can structure a team’s strategic planning, and how the critical meso level functions as a connector between individual practice and institutional direction. I am also interested in the question of whether the sixteen positions in the matrix can be empirically validated through observation of change initiatives - and whether the framework’s predictions about mode-level mismatch align with reported experiences of frustration and effectiveness in institutional change work.

If you work in educational development, learning technology, or academic leadership and any of this resonates - or provokes useful disagreement - I would be glad to hear from you.

References

Arnstein, S. R. (1969). A ladder of citizen participation. Journal of the American Institute of Planners, 35(4), 216–224. https://doi.org/10.1080/01944366908977225 — Open access PDF

Bronfenbrenner, U. (1979). The ecology of human development: Experiments by nature and design. Harvard University Press. Harvard University Press — Google Books

Covey, S. R. (1989). The 7 habits of highly effective people: Powerful lessons in personal change. Google Books

Eisenhardt, K. M. (1985). Control: Organizational and economic approaches. Management Science, 31(2), 134–149. https://doi.org/10.1287/mnsc.31.2.134 — Open access PDF (Illinois)

Fairman, J. C., & Mackenzie, S. V. (2012). Spheres of teacher leadership action for learning. Professional Development in Education, 38(2), 229–246. https://doi.org/10.1080/19415257.2012.657865 — Open access PDF (UMaine)

Harris, A. (2008). Distributed leadership: According to the evidence. Journal of Educational Administration, 46(2), 172–188. https://doi.org/10.1108/09578230810863253

Harris, A. (2022). Distributed leadership: Taking a retrospective and contemporary view of the evidence base. School Leadership & Management. https://doi.org/10.1080/13632434.2022.2109620

Kenny, N., Eaton, S.E. (2022). Academic Integrity Through a SoTL Lens and 4M Framework: An Institutional Self-Study. In: Eaton, S.E., Christensen Hughes, J. (eds) Academic Integrity in Canada. Ethics and Integrity in Educational Contexts, vol 1. Springer, Cham. https://doi.org/10.1007/978-3-030-83255-1_30

Kezar, A. J., & Holcombe, E. M. (2017). Shared leadership in higher education: Important lessons from research and practice. American Council on Education. Open access PDF (ACE)

Lave, J., & Wenger, E. (1991). Situated learning: Legitimate peripheral participation. Cambridge University Press. https://doi.org/10.1017/CBO9780511815355

Moore, J. L., Felten, P., & Strickland, M. (2018). Developmental pathways and approaches to the Scholarship of Teaching and Learning. In N. L. Chick (Ed.), SoTL in action: Illuminating critical moments of practice (pp. 163–173). Stylus. Stylus Publishing

Ouchi, W. G. (1979). A conceptual framework for the design of organizational control mechanisms. Management Science, 25(9), 833–848. https://doi.org/10.1287/mnsc.25.9.833 — Open access PDF (ResearchGate)

Paulhus, D. L. (1983). Sphere-specific measures of perceived control. Journal of Personality and Social Psychology, 44(6), 1253–1265. https://doi.org/10.1037/0022-3514.44.6.1253

Paulhus, D. L., & Van Selst, M. (1990). The spheres of control scale: 10 yr of research. Personality and Individual Differences, 11(10), 1029–1036. https://doi.org/10.1016/0191-8869(90)90130-J

Roxå, T., & Mårtensson, K. (2009). Significant conversations and significant networks — exploring the backstage of the teaching arena. Studies in Higher Education, 34(5), 547–559. https://doi.org/10.1080/03075070802597200

Simmons, N. (2016). Synthesizing SoTL institutional initiatives toward national impact. New Directions for Teaching and Learning, 2016(146), 95–102. https://doi.org/10.1002/tl.20192

Simmons, N. (2020). The 4M framework as analytic lens for SoTL’s impact: A study of seven scholars. Teaching & Learning Inquiry, 8(1), 76-90. Retrieved from https://ezproxy.lib.ucalgary.ca/login?qurl=https%3A%2F%2Fwww.proquest.com%2Fscholarly-journals%2F4m-framework-as-analytic-lens-sotls-impact-study%2Fdocview%2F2412494752%2Fse-2%3Faccountid%3D9838

Wenger, E. (1998). Communities of practice: Learning, meaning, and identity. Cambridge University Press. https://doi.org/10.1017/CBO9780511803932

York-Barr, J., & Duke, K. (2004). What do we know about teacher leadership? Findings from two decades of scholarship. Review of Educational Research, 74(3), 255–316. https://doi.org/10.3102/00346543074003255

more noodling on AI and creativity

nospam-dnorman@me.com (D'Arcy Norman) — Thu, 19 Mar 2026 12:30:59 +0000

The 2026 Oscars™ happened, and a questionable “AI Production Studio” and “AI Talent Agency’s”¹ “AI Actor” tried to use the occasion to convince The Academy™ that replacing human actors with AI is actually good, and that they just need to fully embrace it in order to unlock their full potential. The argument was presented in the form of a soul-less, unartistic, AI-generated “music video” that was basically autotune cranked to 3000 or something, belting out lyrics that had the emotional impact of a corporate press release. If this is “AI Art”, real artists have nothing to worry about.

Technically, it’s impressive as hell. They’ve got persistence of objects and characters, who appear to have the correct number of fingers. Everything is AI-generated². Artistically, it’s just embarrassing and shows that the producers of the video:

don’t understand artistic creation, and/or
don’t care

The audio sounds so overly autotuned (not autotuned, but profoundly synthetic) that it is offensive. The visuals are maybe more convincing but all have the sheen of ChatGPT-generated imagery. This is probably some kind of milestone. But they’re framing it as “Actors. Just give in. It’s happening!” and, no. It’s clear that they don’t understand artistic creation.

There are also related questions about the provenance of the data that was used to generate “Tilly”. Where did the source materials come from? Was it all used with consent? How would that even be possible? Maybe all I need to know about the ethics of this approach is that Mr. Wonderful is in support of it?

And AI is coming for art directors in video games, too. NVidia’s latest “let’s overwork your incredibly expensive graphics card so you can have autotuned visuals in your video game” initiative (DLSS 5), again, looks impressive as hell. Except, it just kind of overprocesses actual creative art direction into a Gritty Reboot™ of whatever’s displayed on the screen.

Again, just looking at the demo video, it’d be easy to ooh-and-ah at the extra details that magically get inserted into everything. Medieval bricks have unprecedented medieval-ness! Shadows are redone, lighting is reworked. Art direction be damned! More pixels! Line goes up!

AI has been doing similar things to our photos for years now - every photo you take on your iPhone etc. doesn’t just take a picture, it takes maybe a dozen and then uses AI³ to smush them together to make the “best” photo from parts of all of them. Take the exposure data from this frame, the reds from that one, this section is more in focus in this frame, and that section is more in focus in that one. Sometimes, it works. Often, it looks autotuned (HDR! It’s a feature!) and overprocessed. Yes, you can turn it off (or can you even do that anymore?) but it’s pervasive and automatic. Years from now, people will be able to identify photos taken in the early-to-mid (to late?) 2020s because of this AI-generated look. Your role as personal art director has been largely removed.

To be fair, there are some interesting and more-ethical (less-unethical? the jury’s still out…) uses of AI in filmmaking. Batman just sold his AI filmmaking tools startup to Netflix for something like $600M. The tools are production workflow things, rather than digital actors. You’ve shot hours of footage of actors dangling on wires and someone needs to digitally remove the lines, so you can train a custom AI model on the footage and say “hey siri, remove the lines” and it’ll take care of it. Or something. Maybe useful? Unless your IMDB filmography credits are all “Digital Effects Artist”?

Other tools, like Epic’s MetaHuman, are intended to be used as authoring tools - using motion capture and voice acting to bring characters into a modeling environment for use in production.

A video is embedded here, but may not be displayed on a preview or if your browser does not support the video tag.

Even this level of software-driven artistic creation is being pushed to be automated with the use of AI - take a single AI-generated photo (ChatGPT - create a photo of a superhero wearing a cape), feed it through some software and Google’s Nano Banana Pro, and get a fully formed Metahuman model to use in your production.

So. Artistic Direction - the Art in art - is being sidelined. This happens because of the “don’t care” aspect. “Whatever, good enough.” That’s basically the mantra of 2026. We’re all so exhausted by everything at every level and every aspect of our lives. Having the energy to be passionate about Creating Art (or anything else) is getting rarer and is hard won. So, the draw of things that don’t require our energy⁴ just naturally become the way we do things. We scroll. We click. We prompt an AI bot to make a video that we can post so other people can scroll and click.

Most people are (or have become (myself included, being a human in 2026)) so separated from artistic creation that it’s an almost alien concept. AI isn’t the cause of that, but it’s sure a contributing factor. I don’t think we get out of this by making more tools, or banning tools, or focusing on the tools. We need to make it a normal part of our lives to make some art, dammit. To take back that sense of artistic direction rather than just offloading that to companies and vendors and going along for the ride.

And - how would that connect with our work in higher education? What if we re-emphasize creation and authorship? How can we address the “beige-ing of thought” in our courses?

Lots of words to get to the crux of it:

How do we get out of the “whatever, good enough” and “don’t care” societal coping mechanisms that have come to define the ‘20’s? How do we embrace the friction and struggle that make inspired creativity possible in a time where everything feels like constant friction and struggle?

see also https://globalnews.ca/news/11713239/ai-actor-tilly-norwood-tillyverse-eline-van-der-velden-xiocia/ ↩︎
Particle6 said that 18 humans were involved in producing it, but they were AI jockeys rather than a traditional film production crew… ↩︎
It started off as not-AI, using deterministic algorithms to do clever math on photos, which, of course, evolved to “let’s use machine learning”, which, of course, evolved into “let’s use ✨Intelligence✨ to do things” ↩︎
Ironically, requiring little of our personal energy requires A LOT of actual energy. Good thing we have such reliable sources of renewable energy powering everything so we don’t have to worry about that! What’s that? Another war in the Gulf? etc. ↩︎

Prototyping a new kind of online course community platform

nospam-dnorman@me.com (D'Arcy Norman) — Wed, 25 Feb 2026 19:41:40 -0700

On the plus side, your humble protagonist has finally figured out how to break out of “the only thing he blogs about is how he uses Obsidian.” Unfortunately, it’s because I appear to be firmly in the middle of a bout of “the only thing he blogs about is how he vibecodes some half-baked idea into a usable thing.”

A blog post in 2 parts:

The Course Community Application
The part about vibecoding and using frontier genAI tools despite everything

The Course Community Application

After my experiment last week in building a local-LLM “course coach” application that students could use to engage with their course materials, a few things happened:

I realized that this is actually a pretty decent way to prototype these half-baked ideas into a form that can be interrogated beyond a blank whiteboard
Brian Lamb shared a link to his Claude-Coded Open Learning Design Toolkit prototype he developed for TRU’s Open Press (which is absolutely AMAZING work! again, the power of a prototype to go beyond whiteboards)
I came back to an idea I’ve had for a few years now - something that I started to figure out back when I was working on my dissertation - of how I would design a “Course Community” platform. Not an LMS. And, although it was out of reach waaaay back in the olden days of 2023, it might just be feasible now

So.

⌘+space Term
mkdir ~/Documents/Development/course-community
cd ~/Documents/Development/course-community
claude

I want to build a web application that will provide rich community interaction between students and instructors within a course. It’s not just a discussion board replacement - it’s a full community environment where people in a course can communicate, collaborate, co-create, give and receive feedback, and more. I want the application to be relatively portable - you can use PHP and SQLite - and it needs to integrate with Brightspace as an External Learning Tool using the LTI specification (which will also handle authentication and course info etc.). Use what you know of innovative and high-impact teaching and learning practices and develop a working prototype that we can try out in a course.

The “use what you know of innovative and high-impact teaching and learning practices” was intended to connect to the documents that I’ve primed my Claude account with.

And then I wondered what would happen if I asked it to refer to my dissertation when building the application, so I fed it the URL to my dissertation and told it to use the framework as guidance.

Claude chewed for a few minutes, asked for some clarification on how I wanted it to work, and then it produced a PHP application that I could deploy on my shared web hosting server. It needed a couple of tweaks, but it was basically up and running. It took me longer to figure out how to set up the LTI 1.3 connection on our Brightspace test server. But that works now too.

I came back with more ideas to add - what about a peer feedback tool? done. what about a collaborative whiteboard tool? OK. what about a place where students can collaboratively edit and publish documents? sure! what about a student response system? etc.

Notably, there is no gradebook or assignment submission tool. The application is aimed entirely at providing tools to use within the context of a course to foster community and collaboration between students and with their instructor. The LMS can do grades and assignments and quizzes, so there’s no need to duplicate that. And, with the LTI integration, this application can become part of the course within Brightspace.

The application is semi-public, running on my sandbox server. The code is on GitHub and I’ll keep that updated as I add features and fixes.

Some screenshots of the current version:

The Community Feed course homepage

The New Post interface

It's not Jamboard, but it's Jamboard-adjacent

Collaborative document editing and publishing

A peer feedback tool

A Pulse Check student response system

Instructor's Course Overview dashboard

The part about vibecoding and using frontier genAI tools despite everything

I am still profoundly conflicted. Here I am, cranking out an apparently endless stream of vibecoded AI slop tools and applications. Steeped in a tool that was built by stealing and plagiarizing, built by ablating semantic complexity in the relentless pursuit of Attention. And propping up the bubble while melting icebergs to do it.

I think there are 2 main reasons that keep me exploring these frontier models despite my discomfort:

In my leadership role, I absolutely have to understand - to know, and to know, how these LLM and AI platforms work. What it’s like to really use them, not just reading about them or making decisions based on outdated info and assumptions. What’s possible, and what the limits are. Because if I don’t have a deep and current understanding, I can’t help to steer institutional policies and initiatives. I need to have this kind of experience to draw upon so I have credibility when talking with other leaders across the university. If I don’t earn that credibility, we’ll be at the mercy of vendors promising things that sound wonderful, and we won’t have any way to push back against the hype.
superficially and selfishly, it’s fun. I can make stuff that would have taken me weeks/months/years to figure out on my own. I can prototype an idea in a couple of hours, share it for feedback, and then move on to another idea. That’s something that I’ve never been able to do in 30 years of building-and-managing. I’ve built big applications. Heck, I’ve built (among other things) 2 LMSs¹², a learning object repository³, and was a core developer of a groundbreaking website authoring tool for use in universities and museums⁴⁵. I’ve done the heavy lifting. And this feels fundamentally different. This feels more like… play?

Does that justify the theft and plagiarism and polar-bear-drowning? No. Of course not. If I could put the genie back in the bottle, I would. I’m not an “AI is inevitable” cheerleader. AI (in whatever form) isn’t/wasn’t inevitable. It’s also not going to go away. It will shift into something else. Into multiple something elses. Metastasize? Fracture? Some portion of it will continue to be useful, will become more useful. Some portion of it will continue to be a corrosive burden on humanity, will become even moreso.

Prototyping a Brightspace Course Coach Application

nospam-dnorman@me.com (D'Arcy Norman) — Thu, 19 Feb 2026 15:42:55 -0700

Last night while watching the Olympics highlights, I was playing around with Claude Code to see if I could implement something I’ve been thinking of for quite awhile. What if students had an application that connected to the LMS (Brightspace in our case) and pulled all course materials, info, calendar, assignments, discussions etc. into a local database, and what if a local chatbot was able to interact with that database to guide a student as they learn? A socratic agent, coaching them without giving them answers. Prompting them as they engage with the course.

I started off simply, using the Opus 4.6 in the Claude app on my phone (again, watching the Olympics). That produced a specifications document that I then saved into an empty project directory. I switched to my laptop and started a new Claude Code session, prompting it to use that document as a starting point for developing a new application.

In a couple of hours of back-and-forth prompting, a pretty interesting prototype application was produced. I wanted to see if it was possible to make a double-clickable macOS application rather than just a collection of server bits that many students won’t want to worry about. Claude suggested using python for the guts of the application, and wrapping an HTML interface in a native macOS application (without using Electron).

Claude started out by being very overconfident in its abilities - it started building to an API that didn’t exist, and asked me to get an OAuth application set up on a D2L server that wasn’t there. So there were a few rounds of “um… I don’t think that’s a thing” followed by “oh, right. silly me! I just made that part up. OOPS! Would you like me to look up the actual documentation and build something that would work?” (yes. please. and why didn’t you start with that? etc…)

Screenshot of the Dashboard interface

Listing the courses that I'm enrolled in through our Brightspace instance.

Image by D'Arcy Norman

On first launch, the application prompts you to login (to our Test1 instance of Brightspace) and then lists the courses you’re enrolled in. Selecting a course will give you some information about it and you’ll need to run a “sync” process the first time so the application can grab all of the info and files.

The application uses a local LLM running in Ollama. I built it using ministral-3:3b but llama would work (if you can stomach running Meta’s LLM on your computer) or others. Course info is processed and stored in a SQLite database for use by the application.

The Course view

With a way to view course materials, content, etc. and to interact with a chatbot.

Image by D'Arcy Norman

With the info and files stored in the application’s database, you can then chat with an AI bot that has been instructed to work with you as you learn in the course. The full prompt is available for reference. The prompt also includes a section based on Natasha Kenny’s executive coaching work with Claude.

Previous chats are saved

and can be exported as markdown files for use outside of the application

Image by D'Arcy Norman

The idea is that you’d use this as a “thinking out loud” partner as you work through the course(s), and that you’d be able to refer back to the ideas that you had during the conversation. The chatbot isn’t an answerbot. It’s not a cheatbot. It’s a power tool to help you do the work to be creative and to more deeply engage with the course.

Chat history is stored locally

and can be filtered by course

Image by D'Arcy Norman

I’ve put the early rough prototype code in GitHub - it may be useful as a reference at least. It’s basically hard-coded to work with our Test1 Brightspace server, so wouldn’t be directly usable, and it still needs a working Ollama environment to run the LLM component. The application is macOS-only, but the python script can also run as a standalone webserver to run on other platforms.

not another think piece on generative AI

nospam-dnorman@me.com (D'Arcy Norman) — Tue, 17 Feb 2026 15:20:33 +0000

AKA Please, God, Not Another AI Thinkpiece!

Nope. There have been enough of those lately. Recent posts about art, intuition, semantic ablation, cognitive debt, cognitive shortcuts and atrophy. They get at lots of the nuance hidden between “AI is literally SATAN” and “I, for one, welcome our new AI overlords!”. Mostly, (generative)AI is kinda useful for some things, is extremely problematic for many reasons, and isn’t going away no matter how much anyone wants it to.

I’ve been using Claude¹ a lot - Claude-Coding up a storm, building lots of little shiny trinkets and tools. Several of these, I use daily. My most-used GenAI-built thing is my bookmarks application², but there are a few other daily drivers. And I have a whole collection of custom Obsidian plugins³ and standalone scripts and applications that aren’t even listed there. Claude also automatically generates tags, descriptions, and alt-text for all of the photos that I post here. I also use Claude for summarizing things, converting handwritten notes into markdown, etc.

Screenshot of Claude Code in the macOS Terminal application

The stats only shows usage since migrating to our Claude Teams accounts and after Claude enabled Code for all Teams accounts.

Image by D'Arcy Norman

The revelation I’ve had recently is that I use genAI for making things that I don’t “care” about. The web-tools? I care about the thing that is made (sometimes) but never about the code itself. The summaries are for things that I didn’t have the energy to read in the first place and only needed a passing understanding of what a thing was about - not a deep, close reading, but more of a “holy hell this is a 20 page article that should be maybe 3 paragraphs - see what you can pull out of it” kind of thing.

So, for those web-tools, it’s not about the code as much as it’s about just willing something into existence in between (during) meetings to see if it’s useful. Most of them get used a bit and then discarded. A few get used every day. And I didn’t have to stop and learn 2026 web design frameworks and new CSS (and and and) in order to get something that worked. The usefulness was in having a thing, not in the building of it.

And, yes, of course, technical debt and cognitive debt and shortcuts, but also art and intuition and playful creation that wouldn’t have taken place without the tedious work being handled by genAI tools.

I won’t use Claude to create something that I want to actually learn or for something where the entire process is important to me. I think it’s potentially useful to be able to decide that I don’t need to create the universe to make an apple pie, because that means I can reallocate energy ~~to let me care more deeply about the things that are more important to me, and to my work.~~ (DLN: ok. that last bit was overblown. I’m writing about making web trinkets and then trying to connect to Deeply Caring, which, sure, but that’s quite the leap. anyway…)

our team is piloting a “Claude Teams” license for a year to see how it might be used - having access to a genuine “frontier model” is pretty important when we’re trying to have informed discussions with instructors and university leadership about what generative AI can do and what the implications are, and Copilot just isn’t there. ↩︎
A pretty feature-complete custom-built drop-in replacement for Del.icio.us etc, but for one user. My live Bookmarks application is at https://darcynorman.net/bookmarks/ and the source code is at https://github.com/dlnorman/standalone-bookmarks ↩︎
Including an LLM “related notes” plugin that uses vector math to find notes that are similar to the current note without having to rely on explicit tags and links, a “People” and “Topics” interface that make my notes extremely useful and efficient, etc. ↩︎

Experimenting with RSS Reader Interfaces

nospam-dnorman@me.com (D'Arcy Norman) — Thu, 29 Jan 2026 10:33:12 -0700

I’ve been using RSS readers for over 20 years. Most of that time has been spent using the excellent NetNewsWire application, but I’ve used others (including Google Reader, Fever˚, etc.).

After reading Terry Godier’s post on RSS readers being stuck in the email metaphor, I wanted to experiment with some ideas for a “non-email” metaphor for a feed reader interface. The most interesting and useful version of this that I’ve used was the “Hot” view from Shaun Inman’s Fever˚ application. What would it look like to integrate something like that into my NetNewsWire database?

Levine’s Law: Start with the demo.¹ (if my dumb webcam self is in the way, you can just drag it if you want so see something that’s blocked by it…)

The main idea is to have an interface to control a set of weighted parameters that act as indicators of “interestingness” for feeds and items. These weights adjust an algorithm that calculates a Score for each item, and that Score is used to present and prioritize items in the reader interface.

This experimental mockup application was vibecoded with Google’s Antigravity and Claude Code. It’s an electron app. Which, yeah. Let’s not do that. But, as a quick way to prototype an idea? It worked well enough. I did burn through a week’s worth of tokens in both Antigravity and Claude Code…

RSS Feed Item Scoring

A screenshot of an interface to set weights for various attributes of RSS feed items.

Image by D'Arcy Norman

This is all basically riffing on the original Pagerank idea (pre-advertising-and-enshittification), and I think that having the weights of parameters used in the algorithm visible - and adjustable - could make this a much more useful approach.

The documentation of the Score algorithm weights is available (generated by Claude Code after it built the Electron app).

Anyway. A rough prototype of a rough idea of the start of maybe something useful?

Levine’s Law also dates back 20 years! Amazing. ↩︎

another 365 photos

nospam-dnorman@me.com (D'Arcy Norman) — Sun, 18 Jan 2026 20:32:51 -0700

Thanks to gentle nudges from Alan, I’ve wrapped up another 365photos “daily photo for a year” project for 2025 and started one for 2026. I’d marked this on Mastodon, but hadn’t mentioned anything here on my blog. Maybe I was waiting to see if I’d actually go for another round? 18 days into 2026 and it looks like I’m still doing it, so…

I started back in 2007, mostly as a way to learn to use my then-new, then-fancy Canon Rebel XT. I carried it in my backpack everywhere and got some interesting shots. I learned to use it pretty well and learned to see differently. Apparently, something clicked because years later I was able to get into a PhD program that required an “artist’s portfolio” and I got in on the strength of my photography. Who knew? #365photos was the beginning of that.

I did miss several photos in 2018 (cancer diagnosis kind of sucked the energy out of the project (and other things)), 2019 (treatment, aging parent stuff), 2020 (something about a pandemic or something, dad died), and 2021 (still stuck at home through the “tail end” of the pandemic, mom died). I got back on the wagon for 2022 and have (so far) stayed on track.

19 years later, and I haven’t touched the XT for a long time, but I still carry the lessons from those days with me. My main camera may be a phone now (my secondary camera is a tiny drone), but it’s less about the gear and more about the practice anyway.

365Photos

Screenshot of my 365photos page, displaying thumbnail images for each of the 20 years I've been doing this thing.

Image by D'Arcy Norman

There are many (many) days where I feel like I’m just phoning it in and grabbing some dumb photo to check a box. But there are many (many) more days where I’m grateful that I’ve built this as a practice because I’ve captured something that will help me to remember something I’d have otherwise let be forgotten as ephemeral.

Anyway. Here’s to another round. Or 20.

copilotting

nospam-dnorman@me.com (D'Arcy Norman) — Mon, 05 Jan 2026 15:17:54 -0700

Microsoft keeps renaming things to be various versions of “copilot”. A snarky toot pointed out that there were now 4 MS products called Copilot. I thought there were more, so I asked Copilot (MS365 Copilot™). It thinks there may be at least 7 different Copilots? But its LLM may be out of date…

Prompt:

How can I keep track of all of the things that Microsoft now calls “Copilot”? Is there a map or guide?

AI-Generated Content via MS365 Copilot

You’re not alone—“Copilot” now spans a whole family of Microsoft offerings across consumer, enterprise, developer, security, and even hardware (Copilot+ PCs). Here’s a compact, practical way to map the landscape and keep it straight, plus trusted sources to track changes as they happen.

The Copilot “family” at a glance

1) Microsoft Copilot (free, consumer & work account)

Chat-style assistant in Windows, Edge, and the web (copilot.microsoft.com). With a work/school sign‑in, prompts are protected by commercial data protections, but it does not access your tenant data unless you have Microsoft 365 Copilot. ramsac.com

Good for: general answers, drafting, summarizing, image generation. ramsac.com

2) Microsoft 365 Copilot (paid add‑on for orgs)

Enterprise-grade AI woven into Word, Excel, PowerPoint, Outlook, Teams, and Microsoft 365 Copilot Chat that can securely use your Microsoft Graph (emails, files, meetings, chats) for context. Licensing prerequisites apply. learn.microsoft.com, learn.microsoft.com

Why it’s different: in‑app experiences + tenant-grounded chat; admin controls, analytics, and agents.
Licensing & prerequisites: Microsoft 365/Office 365 base plans; add‑on pricing and gov variants documented here. learn.microsoft.com learn.microsoft.com

3) Microsoft 365 Copilot Chat (included with many M365 subscriptions)

A web/app chat experience (m365copilot.com / Microsoft 365 Copilot app) with enterprise protections; complements (but doesn’t replace) the paid Microsoft 365 Copilot in‑app features. learn.microsoft.com

4) Copilot Studio (low-code to build custom copilots/agents)

Create and govern task‑ or process‑specific copilots, connect data, extend M365 Copilot with plugins; licensed per tenant with message capacity. licensings…hool.co.uk, powerplatf…rmtips.com

5) GitHub Copilot (developer)

AI pair programmer in VS Code/Visual Studio; different (and separate) licensing tiers for individual vs enterprise. powerplatf…rmtips.com

6) Security Copilot (security operations)

AI assistant for SecOps workflows (Microsoft Sentinel, Defender, etc.); separate SKU and scope. (Microsoft Learn and product pages group it under “Which Copilot is right for your org”.) learn.microsoft.com

7) Copilot+ PCs (hardware category)

Windows devices with powerful NPUs (≈40+ TOPS) enabling local AI features like Cocreator, enhanced Studio effects, Live Captions/translation, and the evolving Recall experiences—rolling out across Snapdragon/AMD/Intel. learn.microsoft.com, blogs.windows.com, yourwindowsguide.com

Name changes & why it got confusing

The Microsoft 365 app (formerly Office) was rebranded to the Microsoft 365 Copilot app in early 2025, while “Copilot” also remains the name of the AI assistant. This dual use of the brand caused understandable user confusion. windowsforum.com, pcgamer.com
Microsoft’s guidance article “Decide which Copilot is right for you” clarifies distinctions among Microsoft 365 Copilot, Copilot Chat, Security Copilot, and more. Bookmark this—it’s the closest thing to an official “map.” learn.microsoft.com

A simple decision tree (how to pick the right one)

Need tenant‑aware help inside Office apps? → Microsoft 365 Copilot (paid). Check licensing prerequisites first. learn.microsoft.com
Just want a safe, web‑grounded chat with work protections? → Microsoft 365 Copilot Chat (included). learn.microsoft.com
Build process‑specific copilots/agents or extend M365 Copilot? → Copilot Studio. licensings…hool.co.uk, powerplatf…rmtips.com
Developers writing code? → GitHub Copilot (dev SKU). powerplatf…rmtips.com
Security analysts? → Security Copilot. learn.microsoft.com
Evaluating hardware for local AI features? → Copilot+ PCs overview & device guidance. learn.microsoft.com, blogs.windows.com

How to keep track (your ongoing “radar”)

Bookmark Microsoft Learn pages

“Which Copilot is right for your organization” (updated frequently, links out to service descriptions and admin controls). learn.microsoft.com
Microsoft 365 Copilot licensing page (details/prereqs, gov/edu notes, links to service descriptions). learn.microsoft.com
Copilot+ PC training module (hardware/category definitions). learn.microsoft.com

Track licensing & naming updates from reputable partners
Mirazon’s 2026 licensing changes overview and similar partner write‑ups often summarize changes swiftly and pragmatically. mirazon.com
Watch product blogs for feature rollouts
Windows Experience Blog for Copilot+ PC features (e.g., Live Captions, Cocreator) across chip families. blogs.windows.com
Use a one‑page internal “Copilot cheat sheet”
Keep a simple matrix: Product → Scope → Where it runs → Access/License → Good for.

It’s easy!

How many confused instructors, students, and staff will we have, after the announcement that Copilot is our supported AI platform. Which copilot? Just the free one? Just the Windows one? All of them? etc.

on building an Obsidian Related Notes plugin using Google's Antigravity IDE

nospam-dnorman@me.com (D'Arcy Norman) — Fri, 21 Nov 2025 11:13:36 -0700

TL;DR:

I built an AI/LLM-powered “related notes” plugin for Obsidian. It seems to work, but who knows? There’s a video tour, and the code is in a GitHub repository.

I’ve been using Obsidian for a few years now, and have always wanted a good “Related Notes” plugin to help me find things that overlap with what I’m working on - based on the content itself, not metadata or links. Haven’t had any luck so I kind of gave up.

Yesterday, a faculty member nudged me to try Google’s new Antigravity IDE. It’s an AI-powered coding application, with Google’s Gemini baked right in. Kind of like Claude Code, but with an actual interface and maybe more generous limits¹?

I downloaded Antigravity and just gave instructions into the chatbot sidebar. I forget the exact prompt, and it won’t let me scroll back that far², but it was something like:

I want to build a plugin for Obsidian that will use an LLM (locally running in Ollama) to index all of my notes and make suggestions for notes that are related to the current note. The list should be displayed in a sidebar.

It chewed for a bit, created some typescript and other files in a temp working directory - I hadn’t even created a project directory yet - and got started. After some back-and-forth, it told me how to compile the typescript projects into the javascript that would be used by Obsidian.

% npm run build

I then created an obsidian-related-notes folder in my Obsidian vault’s .obsidian/plugins/ directory, turned it on, and hey presto. It wouldn’t run. So I checked the debug log, fed that to Antigravity³/Gemini, and it came up with a fix. Recompile, Re-copy to plugin directory, Disable/reactivate the plugin. (this process was repeated many, many times through troubleshooting and debugging). And it worked. I fed it ideas for improvements, I tested them out, and we wound up with a version that works pretty well.

Obsidian Related Notes (AI) plugin

screenshot of the Obsidian Related Notes (AI) plugin, showing the settings screen and kinda showing the plugin running in the sidebar

Image by D'Arcy Norman

I have no idea how it actually works, though. It’s using nomic-embed-text⁴ to calculate vectors for each note. I know very vaguely what that means⁵. And then it compares the vectors for the current note with the vectors in the data file. Somehow.

On my M2 Macbook Air with 16 GB of RAM, it takes about 30 minutes to run the full indexing process on my 6,302 notes. Incremental updates to the index take seconds. BUT - you need to leave Ollama running for indexing to take place.

A quick video tour of the plugin.

It seems to work. I don’t know if it’s useful, or if I’ll keep using it or keep updating it (or keep telling Antigravity to tell Gemini to update it). There’s no warranty, support, or anything.

I’ve put it in a GitHub repository so it’s available if you want to try it out.

especially considering it’s (currently) free, but who knows? ↩︎
note to self: find a way to automatically record all prompts! ↩︎
for some reason, I keep wanting to call it Antimatter instead of Antigravity. Ominous? ↩︎
https://ollama.com/library/nomic-embed-text ↩︎
I mean. I took linear algebra twice because I had to, not because I could actually do it… ↩︎

UCalgary Conference on Postsecondary Learning and Teaching - Call for Proposals

nospam-dnorman@me.com (D'Arcy Norman) — Fri, 07 Nov 2025 09:18:21 -0700

I have the honour of co-chairing (with Natasha Kenny) our 2026 University of Calgary Conference on Postsecondary Learning and Teaching.

The conference is shaping up to be an absolute banger. The theme is From Disruption to Connection: Digital Transformation in Postsecondary Education - a timely topic, given… gestures wildly. We have some AMAZING keynotes lined up (watch for announcements in early 2026). The entire 3-day conference is going to be an incredible showcase of communities and our collective work as we are all navigating, shaping, creating, fostering, and responding to various layers of digital transformations.

Dr. Eliana El Khoury's keynote at the 2025 conference

Here’s the announcement of the Call for Proposals:

We are thrilled to announce the call for proposals for the Conference on Postsecondary Learning and Teaching. This year’s conference theme is: From Disruption to Connection: Digital Transformation in Postsecondary Education

April 28 – 30, 2026

The 2026 University of Calgary Conference on Postsecondary Learning and Teaching invites faculty, staff, students, and postdoctoral scholars to explore digital transformation in postsecondary education. Whether it’s GenAI, teaching methods, or pedagogical shifts, there is incredible opportunity for postsecondary education to spark innovation and ingenuity in educational practices, communities, and systems. Through the exploration of eleven threads, this fully in-person conference will provide you with a unique opportunity to connect, share, and exchange knowledge across three days of interactive sessions, presentations, and more.

Submit your proposal by Jan. 9, 2026.

We look forward to connecting with you!

I’m looking forward to learning from our community throughout the conference. I’m constantly impressed and amazed at how the conference has evolved since we first hosted it in 2015. I think every aspect of our work has changed at least once over that time, and the 2026 conference will be an opportunity to both step back and reflect on these transformations, and to come together to think about how we can best move forward.

Home onD'Arcy Norman, PhD

Experimental Obsidian 3D Graph Renderer

Screenshot of the Graph Render Obsidian plugin

🤖 AI-Generated Content via Claude Code (Sonnet 4.6)

3D Graph Render — Obsidian Plugin

Features

Rendering — Galaxy Style

🤖 AI-Generated Content via Claude Code (Sonnet 4.6)

Filters (Filter button → right panel)

Search

digital transformation is a teaching and learning opportunity

Linkblogging

How I built the Typeset custom theme for this website

Why build a custom theme?

The Series page

Why ’typeset'?

CSS themes

Content types

Obsidian as CMS application

The typeset theme files

Server-side integrations

How it was built

DICE in Practice: Adopting Generative AI Guidelines

The scenario

Mapping the scenario

Reading the matrix

Using this for your own context

Introducing the DICE Framework for Higher Ed Change Leadership

The Problem

The Four Modes

The Organizational Levels

The 4×4 Matrix

Developmental Pathways

The Relationship to RACI

Using DICE for Planning

Future Directions

References

more noodling on AI and creativity

Prototyping a new kind of online course community platform

The Course Community Application

The Community Feed course homepage

The New Post interface

It's not Jamboard, but it's Jamboard-adjacent

Collaborative document editing and publishing

A peer feedback tool

A Pulse Check student response system

Instructor's Course Overview dashboard

The part about vibecoding and using frontier genAI tools despite everything

Prototyping a Brightspace Course Coach Application

Screenshot of the Dashboard interface

The Course view

Previous chats are saved

Chat history is stored locally

not another think piece on generative AI

Screenshot of Claude Code in the macOS Terminal application

Experimenting with RSS Reader Interfaces

RSS Feed Item Scoring

another 365 photos

365Photos

copilotting

🤖 AI-Generated Content via MS365 Copilot

The Copilot “family” at a glance

1) Microsoft Copilot (free, consumer & work account)

2) Microsoft 365 Copilot (paid add‑on for orgs)

3) Microsoft 365 Copilot Chat (included with many M365 subscriptions)

4) Copilot Studio (low-code to build custom copilots/agents)

5) GitHub Copilot (developer)

6) Security Copilot (security operations)

7) Copilot+ PCs (hardware category)

Name changes & why it got confusing

A simple decision tree (how to pick the right one)

How to keep track (your ongoing “radar”)

on building an Obsidian Related Notes plugin using Google's Antigravity IDE

TL;DR:

Obsidian Related Notes (AI) plugin

UCalgary Conference on Postsecondary Learning and Teaching - Call for Proposals

Dr. Eliana El Khoury's keynote at the 2025 conference

AI-Generated Content via Claude Code (Sonnet 4.6)

AI-Generated Content via Claude Code (Sonnet 4.6)

AI-Generated Content via MS365 Copilot