Plain-language glossary

A–Z

Term	In plain terms
Data safe / pod	A secure, private lockbox for your digital information. You hold the key and decide who looks inside, at what, and for how long.
Databox / secure data vault	An organisation's version of the same idea: a private exchange point it runs for each relationship—with a person, organisation or agent—under declared rules, with signed receipts both ways.
Solid	The open standard—led by the web's inventor, Tim Berners-Lee—that makes data safes possible: how information is stored, who may see it, and how software asks permission.
Web of data / linked data	Information published so pieces of it can connect—like library books that can talk to each other to assemble a complete answer.
Ontology / information map	A shared dictionary that keeps everyone—and every computer—meaning the same thing by the same word.
Knowledge bank cooperative	A community institution that runs the shared digital machinery for its members, the way a credit union runs shared financial machinery—one of a network of local cooperatives, not a single central body. "Knowledge" because it curates the logic over the storage: data is raw bytes, information is organised records, knowledge is the meanings and rules that make information usable and trustworthy.
Principal, agent & sub-agent	The legal chain of command. You are the principal—the only one who can authorise action. A fiduciary agent (the cooperative) is bound like a lawyer to serve only you. Software is a sub-agent—an obedient tool with no rights or motives of its own. Everyone else is a third party: outside the wall unless you say otherwise.
Digital homelessness	Owning no part of the digital spaces your life runs through—accounts, records and social life rented from platforms that can change the rules or the locks at will. It also means the records held about you are rarely kept to a standard that could protect you in a court of law.
Community ground	An upgraded regional site—power, water, amenities, workshops, connectivity—where people priced out of housing can live legally and rebuild. A civics space: run by and for the community, not a government program that parks people out of sight.
A vote	More than a ballot at an election: every meaningful choice a person makes—the path they take, the effort they give. A contribution to a cooperative project is a vote, logged and recognised; and a stable place to live is what gives a person the capacity to choose at all.
The journey out	The point of a community ground: the day a resident no longer needs it—leaving with skills, a contribution portfolio and somewhere to go.
Catalytic infrastructure	Spending that builds permanent capability instead of renting temporary relief—solar, workshops, connectivity, micro-loans for homes on wheels—rather than emergency motels that leave nothing behind.
Islandable microgrid	A local power system—solar and batteries on the ground—that can keep running on its own when the main grid fails: cheaper power day to day, and resilience in a disaster.
Sweat equity	The value of unpaid work on a project, recorded as an obligation cost so it can be recognised and repaid—not a donation that disappears.
Community workspace / maker sheds	Shared working spaces on a ground—tools, benches, desks—where residents repair, study, build and collaborate without paying commercial rent.
Guardianship	A declared duty to act for someone's needs in a defined area—a GP coordinating specialists, a parent for a child, a guardian where capacity is impaired. Scoped, visible and revocable: the person stays the principal wherever they can decide for themselves.
Informatics fiduciary	A person or operator legally bound—like a lawyer or doctor—to act in your interests and never profit from what they hold for you.
Artificial intelligence (AI)	Not one thing—a family of techniques for making software do what used to need a human. The branches worth telling apart: symbolic reasoning (exact rules and logic—checkable, provable), machine learning and deep learning (statistical patterns learned from vast examples—powerful but approximate), generative AI (producing text and images from learned patterns), and perception systems (image recognition, natural-language processing—turning the messy world into data). Different tools, different failure modes.
Reasoning engine	A computer helper that follows strict rules over verified facts and can show its working—unlike a chatbot that guesses. The "symbolic" branch of AI.
Inference	Drawing a conclusion from information. Two meanings worth separating: in logic, deducing what follows from facts and rules (what a reasoning engine does—checkable); in modern AI, running a trained model to produce an answer (what a chatbot does—its best guess).
Generative AI	AI that produces new content—text, images, code—from patterns learned in training. Fluent and useful, but it predicts what sounds right; it does not know what is true. Excellent for drafting and translation; not a source of facts on its own.
Lawful access	Access to information because a court ordered it—the specific facts the order directs, nothing more. Not the same as a licence, which is a blanket permission to use or sell someone's data; local infrastructure meets the first without conceding the second.
Machine learning / deep learning	AI that learns statistical patterns from enormous numbers of examples rather than following written rules. Excellent at perception and prediction; poor at explaining why.
Verifiable credential	A digitally signed statement that proves a fact—an electronic certificate, receipt, warranty or permission slip—readable by people and machines, held by you in your data safe and shown only when you choose. Not an identity card: it supports your agency without fixing a label on you.
Open standard	A published rulebook anyone may use—no licence fee, no permission needed—so different systems can work together.
Choice of law	The contract clause deciding which country's courts hear a dispute. Foreign platforms put yours overseas; local infrastructure keeps it home.
Resource commit	A logged contribution to a cooperative project—time, materials, expertise—recorded so it can be recognised and compensated fairly.
Obligation cost	The recorded value of work someone contributed unpaid. When a project earns, that record is repaid—fair compensation for real work, not royalties forever.
Semantic currency	A formal record of non-money value—character, reliability, mentoring, disaster response—so goodwill is recognised and cannot be quietly exploited.
Community portfolio	A collection of your own recognised contributions—skills shown, hours given, work done—curated by you, held in your data safe, and shared only when and with whom you choose. Not a record kept about you: an asset you hold. (The technical design calls these contribution records.)
Civics (not civic)	"Civic" usually means government—institutions duty-bound to serve the public. "Civics" here means the other thing: people themselves doing work for the betterment of their community. A civics project may later be handed to government as a civic resource—but it begins as the community's own.
Individual agency	The principle that each person controls their own digital life and is never merely a data source for someone else's system.
Edge appliance	A small local computer that runs community services on-site—working even when the wider internet is slow, expensive or down.
Social web	Social functions—relationships, messages, sharing—built into the web's own protocols rather than into a company's platform. Communities run their own social spaces under their own rules, interoperating without a global gatekeeper.
Interoperability	Different systems working together because they follow the same published rules—like trains that can run on any track built to the same gauge. It is why a tool built for one community can run on every community using the same standards.
Namespace	An agreed addressing scheme so different systems find the same thing by the same name—like every street in a region following one map. It is what lets separate communities' systems query each other without a central platform.

Where to go from here

The rest of the site uses these words in context: start here for the overview, the databox page for the exchange machinery, and the review page for the honest list of assumptions and open questions.

Continue exploringThe idea