Part 1 – Identifiers

  • What do content identifiers do?
  • They identify content, based on some information that has been provided
  • Shorthand for an exact description of a particular content
  • They don't make assertions about ownership, rights, how the content can be used, etc
  • They can lead to separate, more complete, metadata and descriptions from other commercial and public sources
  • They should be interoperable with many other identification systems—commercial, international standard, industry standard, public/State, private, etc.
  • Why are AV content identifiers important?
  • Metadata is the language of humans to identify things; identifiers are the language of computer-based systems
  • Without reliable identifiers, you (and your system) don't know easily what content and/or specific asset/video/clip/program/etc your catalog/rights document/cue sheet/sale bill/transfer request/encoding order/etc is referring to
  • They make it completely clear which version of something you're talking about – e.g. the German subtitled or the German dubbed version of the original French release or the edited-for-Germany release
  • They can increase AV workflow efficiency
  • They can increase accuracy of AV measurement, tracking, and reporting
  • Although not a right identifier in itself, they make it easier to identify AV works and, therefore, right holders thereof
  • For example
  • They improve efficiency
  • They make it simpler for rights holders, licensees, and distributors to communicate with each other
  • They reduce the costs of moving digital assets through a complex distribution chain- from creation, to post-production, to offer/retailing, to distributing, to reportingand measurement, all moving towards automation
  • They simplify, automate and reduce errors in usage reporting
  • They improve measurement of uptake
  • They improve the management, research & tracking of things in relation to AV works (including rights).
  • They can be used in relationships to group or join things
  • They can be used in conjunction with other identifiers (musical works, sound recordings, books, text, robust commercial metadata sources, etc.)
  • They allow connecting of information from multiple systems (archive catalogs, commercial metadata, distribution points)
  • What are the requirements of an identifier?
  • Unique number referring to unique content within the scope of the ID
  • Adaptability :
  • Coverage at the right level of granularityfor a wide range of use cases
  • Flexibility and extensibility to deal with new circumstancesand use cases
  • Neutral governance structure that gives a balancedvoice to the various communities it is designed to serve
  • Improves stakeholder buy-in
  • Allows reaction to changes in requirements and uses
  • Includes content creators, registrants, users, technologists, application developers, archives, rights holders, etc.
  • Authoritative :
  • Supervision by recognized authority (e.g. ISOor industry-driven standards body)
  • Data supplied by trusted sources :producers, archives, metadata authorities, and other key industry members.
  • High degree of accuracy, and well-understood mechanisms for reporting and fixing errors
  • Persistence :
  • Persistence and maintenance of the standard’sspecifications
  • Persistence of the registry (of identifiers with their corresponding descriptive metadata)
  • Guaranty that ID relates to the same content for perpetuity
  • Interoperability (because there will never be only one identifier)
  • Explicit support of identifiers from other systems or domains
  • Use in other formal and industry specifications (MPEG, DVB, AACS & Blu-Ray, Digital Cinema, CableLabs VOD, Ultraviolet…)
  • Flexible data formats and registration methods
  • Machine to machine interfaces, APIs, and web services for easy access and interoperability
  • Free use and circulation of the ID itself in order to enable applications and support use cases
  • Accessibility of registry
  • Resolve the ID to its associated defining metadata from the central registry
  • As needed for some uses, access to the registry through appropriate mechanisms (APIs, web services, cache copies, …) to enable workflow automation, integration, and other efficiencies
  • Support for external services (extended metadata, distribution, reporting, etc.)
  • Commercially reasonable/nominal costfor issuance, use, implementation
  • What are they used for?
  • Workflow management,workflow automation, distribution
  • rights management (declaration, reporting, payments, …),rights tracking rights clearance
  • cross-referencing, catalogue matching
  • metadata/information retrieval & aggregation, asset retrieval

Current concern:

2 AV content identifiers aiming at the same purpose

  • ISAN and EIDR are both established AV content identifiers
  • ISAN : International Standard Audiovisual Number (ISO 15907)
  • EIDR : Industry standard based on DOI International Standard (ISO 26324)
  • ISAN and EIDR as identifiers have the same purpose (identifying AV works and their versions)
  • But the organisations & technical infrastructure surrounding the identifiers is currently targeted at the needs and priorities of different ecosystems
  • As a result, today the two identifiers are either used differently, used by different communities of stakeholders, used in parallel … or even not used at all.
  • The two identifiers are today complementary rather than in full competition with each other:
  • Different use cases -- Small overlap in implemented use cases
  • Different user bases -- Small overlap in users
  • Possible outcomes
  • Merge the two into a single system that issues only one ID
  • Unlikely in the short term
  • Too much sunk investment by the systems and their sponsors
  • Users can't change immediately
  • Make both do all things for all people
  • Expensive
  • Requires both to expand into areas that are not currentlytheir main expertise
  • Close interoperability, so that if you get one, you automatically get the other
  • Get both IDs from a single registration, where needed
  • For the users, no risk of making a wrong decision
  • Takes advantage of the strengths of both systems

Status

  • Detailed technical discussions are already underway
  • Non-technical discussions (commercial, governance) have started as well
  • Parallel, but interoperable systems,is not ideal, but in the short term:
  • It would be much better than what we have now (pick one or the other, and worry if you did the right thing)
  • It is achievable in a much shorter timescale than other solutions
  • It leaves several possible future paths open for technology, financials, and governance, as well as a unified singular system

Consequences

  • for new IDs
  • Anyone who gets an ISAN can generate anddiscover an EIDR ID when they need it
  • Anyone who gets an EIDR ID can generate and discoveran ISANwhen they need it
  • For previously existing IDs
  • The two systems have to do a matching and gap-filling process between the two systems
  • We are already doing technical analysis on this
  • As a result, no one has to worry about making a bad choice.
  • You can get both at the same time, or know that if you have gotten one you can also get the other when you need it
  • If you have your own identifiers, include them too, to increase interoperability with systems that are not already part of the two identifiers' ecosystems.

Example: ITV

  • Currently getting ISAN
  • With proposed solution, can get EIDR ID as needed for EIDR-based applications

Example: Studio

  • Currently getting EIDR ID and ISANs through separate processes for different applications
  • Under proposed solution can get ISAN and EIDR IDs simultaneously through a single process as needed

Generalization of Examples:

  • Currently must get ISAN and EIDR IDs through separate processes
  • Under proposal, can get both through single process

Part 2 -- Registries:

Identifier registries

  • Identifiers need a registry backing them up
  • To guarantee uniqueness
  • To store the metadata
  • To provide resolution, lookup, etc
  • ISAN and EIDR both have this
  • The “identifier registry” must be only for identification
  • ISAN and EIDR both agree on this : key condition to remain a useful and effective content identifier
  • More detailed databases can be built based on the identifier, but additional metadata shouldn't be in the “identifier registry”
  • Other metadata belongs in its own specialized registries. Examples:
  • Rights databases
  • Commercial information
  • Databases of music tracks in AV works
  • Extended metadata on cast, crew, etc
  • Still images, reviews, scripts
  • This is OK as long as you have cross-referencing and the ability to automate. Example
  • Broadcaster uses a common ID to build its program guide from multiple sources
  • Rights society uses AV identifier as key for a list of rights holders for the work in a particular territory
  • Distributed databases linked through common and broadly used standard identifiers enable rapid evolution and efficient implementation of new ideas
  • Longer term, in some cases operational efficiencies may be realized if specific Registries are combined into a single registry

Rights registry

  • Centralized systems vs Decentralized (distributed) systems
  • Both require identifiers
  • Various ongoing projects and applications :
  • Global Repertoire Database
  • centralized; mainly musical content, but also aiming at specific AV content such as videoclips
  • Linked Content Coalition (LCC)
  • Federated/distributed
  • RDI (the EU implementation project of LCC)
  • DCE (the UK implementation project of LCC.)
  • Safe Creative registry in Spain (safecreative.org)
  • FRAME (see WG1 presentation)
  • …
  • The rights registry is as much a legal and social problem as it is a technical one. Especially around
  • Authoritativeness and certainty
  • Completeness
  • Conflict management (this is unavoidable)
  • Trust
  • Critical mass of uptake

Part 3 -- For Archives:

  • There is no reason not to get an identifier, even if the rights to the content are not clear
  • Neither ISAN nor EIDR implies ownership or rights of any sort
  • Choosing one does not preclude using the other
  • Having a standard “external” identifier (often used in connection with specific internal identifiers) can make research of all kinds easier
  • Diligent search (for determining rights, orphan status, etc)
  • Creation of cue sheets for included music
  • Linking to other data sources(editorial content, ancillary material)
  • Scholarly citations
  • Assist automation ofother tasks
  • Digital encoding, digital publication/distribution
  • Discoverability, linked data applications
  • Communication across multiple systems (internal and external)
  • Separation of archival and consumer-facing data
  • parental ratings, language identification, marking, fingerprinting, measurement, etc