Visit qri.io


terms used & description links

Userland Terms

These terms are exposed to Qri users.

dataseta set of components with a history
versiona set of components at a point in logical time
componenta purpose-driven data structure

Implementation Terms

These terms describe technologies & systems that compose into qri functionality. Users shouldn't need to understand these terms to interact with qri. Qri developers & implementation authors will work with these terms regularly.

logbookour basis for coordination and collaboration. Tracks the changes that a user makes to their collection of datasets, recording only the minimum amount of information. A linear / append-only record of history. Should be cheap to write and trade, safe to push and pull as often as possible
qfsQri’s FileSystem - another abstraction that wraps filesystems in general. Used to coordinate between your normal, local filesystem, a cafs-style filesystem (including IPFS), and a cloud based filesystem (over http or other protocol)
dsrefa dataset reference - a string like “dustmop/infoaboutcats” that references a dataset in my (or someone else’s) collection, and also a structured representation of that string after it is parsed
fsiFileSystem Integration - a mode of using qri wherein a user has a “working directory” that reflects the editable contents of a dataset. Created using “init” (to make a new dataset) or “checkout” (for an existing dataset). For command-line, commands can be used without explicitly providing a reference to a dataset, as long as the user is in the working directory. Implemented using a hidden “.qri-ref” file in that directory.
initidInitialization Identifier - the new way that we want to refer to datasets. Derived from logbook activity, basically a hash of the first entry in a logbook created when a dataset gets a name first reserved for it. Is stable, will never change if a dataset gets renamed, or if a user rotates their private key, etc.

Upstream Terms

These are terms inherited from technologies qri depends on.

cid"Content IDentifiers" is a hashing technique that embeds additional info about the hash in question
DIDDecentralized identifiers (DIDs) are a new type of identifier that enables verifiable, decentralized digital identity.
IPFSan external dependency that we use as our data storage layer. Immutable, content-addressable, p2p system built-in
JWTJSON Web Tokens are an open, industry standard RFC 7519 method for representing claims securely between two parties.
libp2pthe peer-2-peer networking stack qri is built atop. Libp2p was initially factored out of the IPFS networking stack.
UCANDecentralized authentication built on the exchange of JSON web Tokens (JWTs)