Ten projects, one each
The cohort is ten this year. So there are no groups: you take one project, you carry it for twelve weeks, and it goes into the Living Commons under your own name. Eight of the ten continue a thread the 2025-26 cohort opened. Two are new ground.
- Cohort
- 102025-26 was 32
- Project
- Individualno groups this year
- Chosen
- Week 1Wednesday 23 September 2026
- Presented
- Week 12Wednesday 9 December 2026
- Dossier due
- Friday 8 January 2027to be confirmed
- Deposit
- AT6012 Living Commons
Thirty-two students made eight groups of four. Ten students would make three groups of three, and a group of three carrying a twelve-week technical dossier hides who did what. At this size the honest structure is one project each, with the cohort working as one room: every Wednesday's third hour is a shared seminar where two projects are on the table and the other eight read them. You are not working alone. You are working in public.
What every project has to do
Six obligations, the same for all ten. They are what turns a subject into a technical argument, and they are how the ten projects stay comparable when the subjects are not.
- Situate it at a scale. Name the scale band you are working at, from material to building to district to territory. Then say what changes if you move one band up and one band down. A claim that survives only at one scale is a claim about that scale.
- Account for it as stock, process and flow. What is held, what transforms it, what moves through. Then put a number on it using Odum's emergy accounting, which measures the available energy of one kind used up, directly and indirectly, to make the thing.
- Atlas it, and not with a site plan. An alternative, inverted or archaeological cartography. Draw what the standard drawing leaves out. This is where the 3D scanning, the LiDAR and the photogrammetry go.
- Instrument it or twin it. Either a digital twin built from a BIM model, or real measurement from sensing, or both with the disagreement between them made visible. A prediction and a reading are two different kinds of number.
- Say what should be unbuilt. What comes down, what stays, what gets added, and the account that justifies each. Unbuilding is a technical proposition with an energy bill, and it is the hardest of the six to defend.
- Critique the construct you inherited. Read the 2025-26 work on your thread, name the construct it was built on, its settings and its assumptions, and say what holds and what does not. Then propose the improvement or the modification. Finding an error in inherited work and correcting it is the point of a commons, and it is marked.
The ten
Eight continue a 2025-26 thread. Two open new ground. Each names the lens it leads with and the platform surface it works on, because you write your findings back into that surface, where the next reader will find them.
| # | Project | Leads with | Works on |
|---|---|---|---|
| 1 | Urban Reclamation. Vegetation returning to paused and abandoned sites. 2025-26 read dereliction as a transitional ecological interval and stopped there. The question left open is which interval, and whether the site is in release or in reorganisation. Inherits: Urban Reclamation, 2025-26 | Complex adaptive systems. The adaptive cycle, located on a real site | Emergy Explorer |
| 2 | Reuse and Re-habitation. Dereliction as a record of the forces that produced it. 2025-26 read decay as storytelling. This year it has to carry an energy bill and a construction sequence. Inherits: Reuse and Re-habitation, 2025-26 | Adaptive reuse and retrofit. The technical case for keeping | AECO2 BIM Canvas |
| 3 | Technofossils. Which materials and technologies persist as geological evidence. 2025-26 catalogued candidates. The open question is residence time, and how you would measure it. Inherits: Technofossils, 2025-26 | Deep time and material stock. Stock before flow | Emergy Explorer |
| 4 | AI-Driven Housing Analysis. Machine reading of housing stock, and what it sees and misses. 2025-26 established that it misses things. This year: which things, how often, and what the error costs a decision. Inherits: AI-Driven Housing Analysis, 2025-26 | Simulation, prediction and its error bars | infrared.city |
| 5 | Auto Amelioration. Systems that improve their own condition with nobody managing them. 2025-26 named the phenomenon. This year it needs a feedback loop drawn, instrumented and measured. Inherits: Auto Amelioration, 2025-26 | Regenerative aspiration. Feedback that runs without a manager | Digital Twin and Sensing |
| 6 | Fluid Memory. Water as a carrier of urban narrative and of physical record. The strongest 2025-26 thread for the atlassing obligation, and the weakest on measurement. Inherits: Fluid Memory, 2025-26 | Inverted and archaeological cartography | The field-trip atlas |
| 7 | Infrastructures of Subjugation. Infrastructure read as a distribution of power across a city. 2025-26 made the reading. This year it gets played, with the people the infrastructure acts on at the table. Inherits: Infrastructures of Subjugation, 2025-26 | Actors and community engagement. Who is in the room | simSPACE |
| 8 | Waterscape. Water ecologies at the scale of the territory. The thread that most obviously breaks when you change scale, which makes it the test case for obligation 1. Inherits: Waterscape, 2025-26 | Scale bands, and what translates between them | Datum |
| 9 | Unbuild. New this year. What comes down, and the account that justifies taking it down. The term has no settled definition in the literature and none in this platform's ontology, so part of the project is proposing one and defending it. New ground. Grounding a term the estate does not yet hold | Unbuild. Dynamic life-cycle assessment, resolved over time | AECO2 BIM Canvas and Emergy Explorer |
| 10 | Playable Ground. New this year. A geogame as a research instrument for urban design and community engagement, where the people affected help design the mechanics. New ground. Participatory game mechanics as method | Games as a way of collecting spatial reasoning from non-specialists | simSPACE and Wellbeing and Occupancy |
The bridge to the Living Commons
Obligation 6 is the one that makes this a commons, so it is worth saying exactly what it asks.
You inherit a thread. Before you extend it you take it apart. The construct: what model of the world was the 2025-26 work built on, and was it ever stated? The settings: what scale, what time window, what boundary was drawn around the subject, and what did that boundary exclude? The improvements: what would you change to make the claim testable? The modifications: what do you actually change, and what happens to the finding when you do?
Write that as a section of your repository entry, in plain words, naming the previous cohort's work. Say where it was right. A critique that finds nothing right is a critique that has not been done carefully.
Then leave the same thing for the next cohort: what you could not settle, what you would do with another twelve weeks, and where you think you are wrong.
Every project above works on a live surface of the ENGAGE platform and writes its findings back into it. That platform is also the thing being studied: how it is built is on the syllabus, at DevOps. You are using an instrument, criticising it, and improving it, in the same twelve weeks. When the instrument is wrong, saying so is part of the work.
What you hand in
Continuous assessment, 100 marks. Students participate in seminars and produce seminar paper and technical dossier. The module delivers that in two individual components of equal weight.
Repository entry and presentation, 50 per cent
- Six to eight pages of designed digital content, roughly 2,000 to 2,500 words.
- Several media working together: text, diagrams, scan or survey output, data visualisations, interactive elements.
- Live links to every reference, precedent and resource used.
- The critique of the inherited construct, per obligation 6.
- A fifteen-minute presentation in week 12, with questions from the other nine.
Marked in four equal quarters: conceptual rigour and theoretical grounding; design quality and the experience of using it; technical implementation; and how it sits inside the repository as a whole. AI fluency is assessed across all four.
Technical dossier, 50 per cent
- The six obligations, each answered and each showing its working.
- The process and the technical decisions, with the reasoning and the dead ends.
- Documentation someone else could work from without asking you a question.
- A 1,500-word reflection on the method, replacing the group reflection that no longer applies.
- An annotated bibliography of at least fifteen sources, APA 7th edition.
- A development pathway for the next cohort.
Every reference in your dossier is checked against the publication record before it goes in. A DOI is read off the publisher page or the Crossref record, or it is left out. AI use is disclosed: which tool and version, what you asked it, how you changed the output, which decisions were yours, and what it got wrong. That is the Diligence competency of the AI Fluency Framework, applied to your own work.
Project brief for 2026-27, written 24 August 2026. Supersedes the 2025-26 group brief, which was built for a cohort of thirty-two. Student names and identifiers have been withheld from the 2025-26 material and are restored for anyone on the consent register. Assessment weightings follow module brief version 3.1. The dossier date is subject to the UCC assessment calendar. Module brief · The twelve weeks · Living Commons.