Glossary

This site uses the vocabulary of two worlds at once: audiovisual production, and data spaces. Neither is obvious from the outside. Everything below is defined in plain terms, and linked from wherever it appears on the site.

Filter

Data space vocabulary

Connector: The software component a participant runs, or has run for it, to exchange data inside the data space. It enforces identity, contracts and policies.; Your connector is your way into the data space, and it works for you. It proves who you are to the others, it publishes what you are willing to offer, and it applies the terms you set on every exchange. You can run it yourself, or have a provider run it for you: either way it is yours, and the rules it applies are the same for everyone.
Contract and policy: The machine-readable terms attached to an exchange: who may use what, for what purpose, for how long.; The terms travel with the data rather than sitting in a signed document nobody reads again. Because they are machine-readable, the connector can enforce them, which is what makes selective sharing more than a promise.
Credential verification: The trusted check that a participant's claims about itself are true, before it is accepted into the data space.; Anyone can claim to be a broadcaster. Verification is what makes the claim checkable by others, without each participant having to run its own investigation.
Data offering: A described, contractable unit of data or service published to a participant's local catalogue.; An offering is what a consumer can actually ask for: a described set of data, with the terms that come with it. Describing it is most of the work, because it is what makes the offering findable and comparable.
Data sovereignty: Keeping the ability to decide who uses your data, how and for how long, even after you have shared it.; Sharing usually means losing control: once a file is sent, it is gone. Sovereignty is the attempt to keep the decision with the owner after the exchange, through terms that remain attached to the data.
Data space: A decentralised infrastructure enabling trustworthy data exchanges based on commonly agreed rules.; A data space is not a platform and not a database. It is the combination of a clearly identified group of participants, a trustworthy governance framework and a technical infrastructure that ensures the rules are followed. The technical components exist to enforce that rules. Nothing is centralised, and each participant keeps its own data where it already is.
Federated catalogue: The mechanism that makes the offerings of every participant discoverable together, without centralising any data.; Each participant describes what it is willing to offer in its own local catalogue; the federated catalogue makes those descriptions searchable across participants. The data itself never moves into it, which is exactly the point: you can find out that something exists without anyone handing it over first. It is not the catalogue of works of a broadcaster, a distributor or an archive: that is a business collection, and it stays where it is.
Local catalogue: The list of offerings a participant publishes from its own connector, describing the data and services it exposes.; The local catalogue is where a participant says what it is willing to offer, and on what terms. It stays with the participant, on its own connector; only the descriptions become discoverable, through the federated catalogue. It is not the catalogue of works of a broadcaster, a distributor or an archive: describing an offering is a deliberate act, not the publication of everything you hold.
Onboarding: The process of admitting a participant: verifying its identity, issuing its credentials, setting up its connector.; Onboarding happens once and conditions everything after it. It is where the trust that later exchanges rely on is established, which is why it is deliberate rather than instant.
Participant: An organisation admitted to the data space, identified under its rules and able to exchange within it.; Becoming a participant means being verified once, then being recognised in every exchange that follows. It is what makes an exchange attributable: each side knows who the other is, and can prove it later.
Participant identifier: The verifiable identity a participant uses inside the data space, so that every exchange can be attributed.; Without a verifiable identity, an exchange is just a file arriving from somewhere. The identifier is what turns it into an exchange between two known organisations, which is the precondition for any contract to mean anything.
Sandbox: An isolated environment for testing exchanges before running them for real.; A sandbox lets a participant get the plumbing wrong without consequences. It matters most for the first exchange, which is the one nobody has done before.

Audiovisual metadata

Asset: A file that belongs to a work: the master, but also a subtitle file, an audio description, a sign-language version, timestamps, etc.; The work is an idea; the assets are the files that carry it. A single work can have dozens, and most of the value that goes unused sits in the assets nobody knows exist.
Descriptive metadata: The fields that say what the work is: synopsis, genre, tags, cast and crew, awards, storyboard.; Descriptive metadata is what makes a work findable by someone who does not already know it exists. It is also the part that is most often rewritten, because everyone describes in their own words.
Identification metadata: The fields that establish which work this is: identifiers, titles, type, year of reference.; Identification answers a single question: is this the same work? It sounds trivial until two databases disagree on a title, and nobody can tell whether they hold one work or two.
ISAN: International Standard Audiovisual Number: the ISO identifier of an audiovisual work, which stays the same wherever the work travels.; An ISAN identifies the work, not a file or a copy of it. Because it does not change from one system to the next, it can act as the pivot that lets two databases agree they are talking about the same thing.
ISRC: International Standard Recording Code: the equivalent standard identifier for sound recordings.; ISRC does for a recording what ISAN does for an audiovisual work. Both matter here for the same reason: a shared identifier is what makes matching possible at all.
Localisation asset: An asset produced so a work can be consumed in another market or by another audience: subtitles, dubbing, audio description, sign language.; These assets are expensive to produce and are usually produced more than once, because there is no way to find out that they already exist. Making them findable is a use case in its own right.
Metadata: The data that describes a work: what it is, who made it, how it is encoded, who holds rights on it.; Metadata is everything about a work except the work itself. It is what makes a film findable, payable and usable, and it is what gets rebuilt from scratch at every handoff when nothing carries it.
Provenance: For each field, who wrote it and when, kept alongside the value itself.; Provenance is what lets you trust a field you did not write. Without it, a correction and a mistake look identical, and the safest thing to do is re-enter everything, which is exactly the behaviour it removes.
Quality control (QC): The checks run on a file to confirm it meets a technical specification, and the results of those checks.; Quality control is expensive to run and cheap to lose: the report often stays in the tool that produced it. Carrying the result with the work means the next participant does not have to run it again.; Also: qc
Record: The whole set of metadata held about one audiovisual work, kept as a single living object instead of scattered across spreadsheets and systems.; A record is not a file and not a form. It is the accumulated knowledge about one work, growing as the work travels, with each field keeping who wrote it and when. Today that knowledge exists, but split across a dozen systems that never talk; the record is the idea of holding it in one place that everyone can read from.; Also: records
Technical metadata: The fields that say how the work exists as a file: resolution, aspect ratio, colour space, codecs, and the results of quality control.; Technical metadata is what a system needs to accept a file without opening it. When it is missing or wrong, the file is rejected late, usually by the person least able to fix it.
Version: A distinct edit or configuration of the same work, which carries its own technical metadata.; A broadcast cut, a festival cut and a streaming master are the same work and three different versions. Conflating them is how a delivery ends up technically valid and editorially wrong.
Vocabulary: The agreed list of field names and terms participants use, so that the same thing is called the same thing everywhere.; Most exchange failures are not technical: two sides use the same word for different things, or different words for the same thing. A shared vocabulary is the least glamorous and most load-bearing part of interoperability.

Actors and rights

Broadcast data: The record of what was aired, when, and on which channel, used to identify works and pay the rights on them.; Broadcast data is produced as a by-product of running a channel, and consumed by organisations that had no part in producing it. That distance is where identification problems come from.
Broadcaster: The organisation that programmes and airs a work, and needs metadata its systems can ingest as-is.; A broadcaster receives from many producers, each with their own habits. What it needs is not more metadata but metadata in a shape its systems accept without a human in between.
Collective management organisation (CMO): A body that collects and distributes rights revenue on behalf of rights holders.; A collective management organisation pays people from data it did not produce. Everything depends on being able to tell which work was aired, which is why inconsistent broadcast logs are not a clerical problem but a payment problem.; Also: cmo · collective management organisation
Producer: The organisation that makes a work, and originates most of its metadata.; The producer is where nearly all metadata is first created, often in a spreadsheet, and where re-keying is most costly to avoid because nothing downstream can invent it.
Registry: An organisation that assigns and maintains identifiers, and vouches for the metadata attached to them.; A registry is what makes an identifier worth trusting: it is accountable for the mapping between an identifier and a work, and for keeping it stable over time.
Rights holder: Whoever holds rights on a work, or on a contribution to it.; A single work can have dozens of rights holders, from the production company to an individual composer. Identifying them correctly is what turns broadcast data into money reaching the right people.
Rights waterfall: The ordered rules by which the revenue from a work is split among its rights holders.; The waterfall says who gets paid, in what order, and up to what threshold. It usually lives in contracts written in prose, which is why extracting it as data is a use case of its own.
Secondary market: The reuse or resale of a work, or of its assets, after its first exploitation.; Most works have a long life after their first broadcast, and most of the value in that life goes unrealised for want of knowing what exists and who holds it.
Service provider: Any organisation offering a service around the data, or the content, of audiovisual works to participants of the data space.; The definition is deliberately wide. Limiting it to connector vendors would describe the plumbing and miss the point: what participants need is access to services, of which connectors are one kind.

TAMIS and Europe

Common European Data Space: One of the sectoral data spaces set out by the EU data strategy. The one for media is TEMS.; The strategy sets out one data space per sector, on the assumption that health, mobility and media do not share the same constraints. TEMS is the media one, and TAMIS works inside it.
Digital Europe Programme: The European Union funding programme co-funding TEMS.; The programme funds infrastructure that the market would not build on its own, which is a fair description of a shared metadata space: everyone benefits from it, nobody would own it.
European Accessibility Act: The European rules that drive demand for subtitles, audio description and sign language on audiovisual content.; The Act turns accessibility from a good intention into an obligation, which is why localisation assets are being produced in volume, and why finding the ones that already exist matters economically as well as ethically.
Gaia-X: The European initiative whose trust and interoperability principles data spaces build on.; Gaia-X did not build a platform; it defined what makes a data space trustworthy. TAMIS inherits those principles rather than reinventing them.
RGAA: The French accessibility standard for digital services, aligned with WCAG.; RGAA is the reference this site is evaluated against. It requires publishing a conformity rate, which is what makes the claim checkable rather than declarative.
RGESN: The French reference framework for the eco-design of digital services.; RGESN sets out 78 criteria across nine themes. It is not mandatory for TAMIS, and the site follows it anyway, publishing what it does not meet alongside what it does.
TAMIS: TAMIS is the trusted European data space where audiovisual metadata stays intact, from production to broadcast to reuse.; It lets the metadata of an audiovisual work circulate between everyone who produces or consumes it, without anyone giving up control of their data. The project was originally named, in French, Traçabilité dans l'Audiovisuel des Métadonnées de manière dIstribuée et Sécurisée, and an earlier English wording spoke of a Trusted Audiovisual Metadata Interconnection Space. Both are history, kept because a name has one: the brand baseline is, and remains, AUDIOVISUAL METADATA SPACE.
TEMS: Trusted European Media Data Space: the European project bringing together over 40 organisations across Europe, of which TAMIS is the audiovisual production trial.; TEMS is the wider project; TAMIS is one of its trials, focused on collaboration in the audiovisual production chain. What TAMIS builds is meant to work inside that larger frame rather than beside it.; Source
Trial: In TEMS, a use case tested in real conditions by a group of partners. TAMIS is Trial 5, on collaboration in the production chain.; A trial is not a pilot in the marketing sense: partners exchange real data, with real constraints, and the result is allowed to be that something does not work.