LCC and the “Digital Identifier Network”€¦ · URI2 URI7 URI1 URI20 URI4 URI8 URI21 URI14 URI19 URI16 URI17 URI18 URI9 URI10 URI11 URI12 URI13 URI15 What this looks like (ideally)

© LCC/Rightscom 2013 Godfrey Rust Licenses for Europe March 2013

LCC and the “Digital Identifier Network”

Godfrey Rust, Rightscom/Linked Content Coalition

Presentation to the “Licenses for Europe – WG2: User Generated Content”

April 2013


Context: the digital content explosion

The number of new and adapted works of all media types

(text, image, audio, audivisual) now loaded or created on

the internet each day is greater than the total number of

books, audio and audiovisual recordings published in

physical form in the history of civilisation.

This figure was negligible ten years ago. The result is an

unimaginable number of new daily digital orphans –

content whose identity and rights are inaccessible to users

or service providers in any automatable way.


This presentation…

Summary of LCC and the RDI project

Brief description of the Digital Identifier Network which as

the context of rights management for the future (using the

LCC’s rights model to describe it)

Suggestion of Web Content Declaration of interest – an

approach LCC is considering for addressing the digital orphan

issue.


Linked Content Coalition (LCC)

LCC was established in 2012 to develop building blocks for the expression and management of rights and licensing across all content and media types.

Membership global: all media types and all parts of the digital content supply chain.

Supported by EC (funding RDI project) and UK “Copyright Works” report leading to Copyright Hub.

Phase 1 completed – “LCC Framework” published last week.

Phase 2 in planning.

LCC to be a consortium of standards bodies?

4


LCC Vision

Use technology to benefit media supply chain participants, not to their detriment.

5


LCC Role

To be a catalyst to encourage the automated management of content rights in the digital network.

6


An LCC assumption

Rights data management is broadly the same in all media –differences of emphasis, not of fundamentals.

7


LCC is not…

…advocating automation where it isn’t appropriate

…about replacing existing standards

…about copyright law

…biased to any sector or business model

8


Potential benefits

Easier discovery of rights ownership will increase market size for

rightsholders and decrease infringement.

Increasing automation will reduce cost and increase profitability for

all supply chain participants.

More standardisation will lower system development costs,

encouraging transformative innovation and increasing market size

for all supply chain participants.

9


First deliverable - LCC Framework (April 2013)

Specification for best practice and interoperability in the digital rights data supply chain.

Specs for Identifiers and Messages (and, in future, user interface/iconography) in the digital network.

Rights Reference Model (RRM) – comprehensive data model for all types of rights in all types of content for all types of use and control.

RRM can be used for system/message design or data transformation for interoperability between other schemas (whether standard or proprietary) – a “hub” model to allow anyone to talk to anyone about rights.

10


RDI (“Rights Data Integration”) project

Beginning (May?) 2013, for two years, EC funded.

An exemplary implementation of the RRM as a “hub” .

Real businesses dealing with real data and hoping for real long-term

business opportunities from RDI.

A range of data flows across the supply chain to show that different

rights expressions (licenses and rightsholding claims) from all media

types can be integrated using an implementation of the RRM.

Also show how new standards can be implemented to fill gaps (in

image sector).

11


mEDRA

RDI participants

12

IFRRO

(Pearson)

Axel Springer

(IPTC)

Publishing Still images

PPL

(EMI Music Publishing)

CI

Music

(FremantleMedia)

Danish Producers

Association

AV

Cineca

Rightscom

NTUA

Ediser

Sources

Transformer/ mapping

Exchanges

(Kobalt Music)

CEPIC Getty

age fotostock/THP

Picscou

t Album

Rights

Direct

Users Brackets denote contributing partner which is not a member of the consortium

PLUS (Capture)/ (BL)

Album


UK Copyright Hub

LCC endorsed by Hooper “Copyright Works Report”.

Hub working groups reviewing LCC Rights Reference Model with a view

to recommending it as the data architecture for the Hub.

13


The Digital Identifier Network (“DIN”)

The rest of this talk will use the RRM to present a vision

of the present and future of the rights supply chain as a

network of digitally resolvable identifiers.


The LCC Rights Reference Model (RRM)

A comprehensive data model for all media and right

types. Based on much previous work and best data

modelling practise.

Designed to cover scope of all known existing rights

standards and more. Extensible, flexible, optimizable.

Tested with use cases.

Data modelling is not rocket science: it is about

describing the reality you want your system to deal

with. A model should make sense to anyone who has

some understanding of that reality. If it doesn’t, your

system will cause you trouble.


The RRM demystifies rights data management

All multi-media rights data, however complex, can be aggregated

and expressed in common and relatively simple ways.

Across-the-board interoperability is achievable: “our domain is

special” is not true.

Differences can all be accommodated by identifier and vocabulary

management.


0-n 1-n

1

1

0-n

0-n 0-n

Link

LCC Entity Model: Attribute Model

Entity

Category

Entity

LCC Entity Model This diagram shows the common structure for each Entity in the RRM (and other models which LCC may specify in future). Each Entity is built in a modular way from combinations of five types of Attribute, each of which has a different “micro-model” structure, exemplified here. Each Attribute is an Entity in its own right and may have Attributes of its own.

Type Value

example xyz:RightType xyz:Play

Type Value

NameType Designation

Part Part

example lcc:Name “John Smith” ReferenceName “Smith, John”, Indexed “John”, NamesBeforeKeyName “Smith”, KeyName

Type Mode

Proximity Value

Unit

example xyz:FileSize lcc:SingleQuantity lcc:NotMoreThan 10 xyz:MB

Descriptor

A Name, Identifier or Annotation of an Entity in the form of an uncontrolled or partially controlled data value

Type Mode

Proximity From

Proximity To

example xyz:ValidPeriod lcc:Period lcc:Exactly 2012-01-01 Lcc:NotAfter 2013-12-31

LinkType Entity1

Entity1Role Entity2

Entity2Role

example lcc:Creation_Party A123 (=“Moby Dick”) B987 (=“Herman Melville”) xyz:Author

Note: Some element names are abbreviated because of space

Category

A categorization of of an Entity with a fully controlled data value

Quantity

A measure of some aspect of an Entity

Time

A point or period of time associated with an Entity

Link

A typed relationship between two Entities


RRM EntityTypes and principal Links

Right

Creation Party

Rights Conflict

Assertion

Rights Assignment

Rightsholder

Controlled Creation

Contributor

Assigner

AssignedRight

SubjectOfAssertion

Asserter

Context

Condition

ExcludedRight

SourceRight

Place

ValidPlace

Assignee

ConflictedRight

ValidContext

This diagram shows the eight EntityTypes of the RRM, and the main links between them.

RRM Entities and principal Links What the reality looks like


ID5

ID6 ID3

ID2

ID7

ID1 ID4

ID8

Type1

Type2

Type4

Type3

Type6 Type7 Type8

Type9

Type10

Type11

Type12

Type13

Type5

What this looks like to a computer

RRM Entities and principal Links Each of the Entities is an Identifier and each of the Links is defined by a Type (a term from a controlled vocabulary). The computer has no idea what it means. It only works if IDs and Types are in a form that it can process. It can recognise if an ID or a Type is the same as another – that is how it is able to pass information from one model to another.

Type14


URI5

URI6 URI3

URI2

URI7

URI1 URI4

URI8

URI20

URI21

URI14

URI19

URI16 URI17 URI18

URI9

URI10

URI11

URI12

URI13

URI15

What this looks like (ideally) on the Web

The Linked Identifier Network On the Web, each ID and Type is best represented as a URI. That enables connections to be made using standard Web protocol. This is creating the Linked Identifier Network on which digital commerce relies. Rights, money, content, searches, piracy and anti-piracy actions all flow along this network.

URI22


URI5

URI6 URI3

URI4

URI8

URI20

URI14

URI17 URI10

URI13

What this looks like (actually) on the Web

The Linked Identifier Network Where there are gaps in the network, the Network is broken.

URI22


The network is broken

The Linked Identifier Network is seriously broken

(or rather, not yet built) in many places, and some

of these get in the way of good content and rights

management.

Either

- IDs and Types are missing, or

- IDs and Types are not mapped to each other

where they represent the same things, or

- they are not yet usable as URIs

That sums up what LCC wants to see done.


Provisos

Not all links are public. The network will always be

confidential or inaccessible in some places for

good reasons.

Not all parts of the network are worth the cost of

building and maintaining them.


Two key developments which shape the solution

• Digital content identification

• “Direct-to-web” publishing.


Digital content identification

There is plenty of technology now for identifying digital

items, for example:

• Digimarc Guardian (Attributor) - text

• Getty PicScout – images

• Soundmouse - audio

• YouTube Content ID - audio visual

These mean that content items can be tracked for any

purpose.


“Mapped”, not just standard, identifiers

The same identifier standards do not need to be used

everywhere – different standards can be “mapped” to one

another so that systems can translate automatically.

This is the basis of the ISNI Party identifier currently being

implemented.

This is also how different controlled vocabularies can be

interoperated (RDI will greatly expand the Vocabulary

Mapping Framework).


“Direct-to-web” publishing

Most items of content are now published first in digital form on

the web. Much of this content is of some commercial value (or

potential commercial value).

Much is “self-published” by individuals or by organizations who

are not primarily publishers but Primary publishers also now

release a great deal of content through websites, blogs and file-

sharing services.

There may now be more than a billion “web content

publishers”.


The identifier gap

Most direct-to-web publishing captures metadata –

who is the creator / rightsholder

what is the content

and sometimes

what rights are available

but much of it is not machine-interpretable (that is – not

captured as shared identifiers) or authorized. It is captured as

free text or proprietary IDs. Unlike data in a public registry, for

example, it cannot


Simply stated…

…most web content publishers have no way of declaring

who they are

what their content it

what content it is derived from (if any)

and

what rights they control or grant

in a form that is authorized, and can be discovered automatically

by anyone who wants or needs to know.


Digital content IDs are dumb

Digital content IDs enable a high level of automated

identification – but they only recognise the content.

Digital content IDs must be linked at some point to “human”

authorized party, content and rights identification.

This set of identifiers might be called the authorized key

metadata.


Declaration “at point of first publication”

The obvious solution is to issue the authorized key metadata

at the point at which a digital item first enters the web.


“Web Content Declaration of interest” (WCD)

This could be a standardised but flexible declaration of

interest of any kind (contributor or rightsholder) in an item of

digital content published on the web.

The WCD would include:

• a digital content ID (automatically generated)

• party ID(s) for the interested party(s)

• creation ID(s) for the content item and linked items

• rights assignments

• other metadata using mapped controlled vocabularie


use case: a music video (but could be any content)

Performer “Ebony Day”

Creation “Kiss You - One Direction (Ebony Day

Live Cover)”

This is a version of: Music video by One

Direction performing Kiss You. (c) 2013 Simco

Limited under exclusive licence to Sony Music

Entertainment UK Limited

Rights No information published.


As a Web Content Declaration of interest

ContentID 0101010101010101010

Performer ISNI 123456789

Creation ISRC 234567890

ddex:MusicalWorkVideo (etc)

xyz:cover of: ISRC 987654321

Rights ABC Rights Assignment 123456789 (etc)

Asserter ISNI 987654321 date 01-01-2015

Most WCD data can be stored as a default “profile” and created automatically as part of the content publication process.


“Web Content Declaration of interest” (WCD)

Based on the LCC models.

Customised for specific content and services using Apps and

APIs.


Some health warnings

Any systems must be voluntary and opt-in.

Not all rights management can be automated.

This approach is not “legal deposit” (though could be

integrated with legal deposit where that exists).


How to get IDs?

Web content publishers need to be able to acquire IDs

automatically.

Existing identifier registries (eg ISRC, ISNI, ISBN, ISWC, EIDR, DOI

etc) can allow web content publishers to acquire IDs directly (a

challenge in volume, cost or authorization for some of them)

and/or

new identifiers and registries will emerge to provide

identification services. These will feed into established registries

where appropriate.


Where to put WCD data?

Content aggregators and service providers would not be

expected to accept any responsibility for the data (unless they

want to).

WCD data can go to existing


“Web Content Declaration of Interest”

Machine-interpretable declarations of content and rights made

at the point of first publication of digital content on the web.

Using (and supplementing) existing identifier and metadata

standards and registries wherever possible.

Declarations to link digital content IDs to authoritative party,

creation and rights identifiers and metadata.

Declarations should be integrated with content loading systems,

but web content aggregators not expected to be registries –

Apps and APIs to provide links.

May be supplemented of course

(user default patterms) (wrks that are registered)

Currently no place to decare rights control


Rights expressions in WCDs

The LCC model supports the simplest or most complex rights

expressions, using any controlled vocabularies.

Other rights languages such as ODRL, RightsML or ONIX-PL can be

used for licenses.

The model can be used to express public domain and orphan works

and any specific provisions or exceptions for specific jurisdictions.

Rights claims can be made at any time after first publication by

association.

Conflicting claims will be a significant issue for WCDs, as they are

for collective rights registries now.

.


Controlled vocabularies

The LCC RRM supports the simplest or most complex rights


kfkfk

.


Impact on licensing

The LCC RRM supports the simplest or most complex rights


Kfkfk

.



“Digital orphans”

If metadata is not machine-interpretable, an item of content

maybe called a digital orphan – it requires human

intervention to complete a chain of transaction.

The problem of new digital orphans is far larger than the

problem of “historical orphans”: billions of digital orphans are

being created daily.


“Rights” in the LCC model

A Right is “a state in which a Party is entitled to do something

in relation to a Creation, as a consequence of a law, agreement

or policy”.

A Rights Assignment is “a decision which results in the

existence of a Right” and so covers any kind of agreement

(including licences), law or policy which grants rights.

Offers and requests can be expressed in the same structure as

“proposed rights assignments”.


Assertions and conflicts

Digital


The missing link(s)

Digital content identification

“Direct-to-web” publishing.






The future of the Digital Identifier Network

Five things I expect to happen, and which LCC will be

trying to facilitate:


The future DIN: Party IDs – first in the chain

Party IDs are the starting point – and the blind spot in

several sectors.

eg CISAC’s great success story – the CAE/IP number.

eg ISNI (and IDs which link to it)

When Creations or Rights are being identified, the

Rightsholders must first be identified if licensing is to be

automated.

All Rightsholders who want them will have IDs (like ISNI,

or an ID mapped to ISNI) which enable them to register

content and rights. They will be able to get them quite

easily.


The future DIN: Creation IDs - “Digital declaration”

Automated content identification is already common in

images (eg PicScout and YouTube) and audio.

Growing use in text world (eg Digimarc Guardian).

Allows a party to declare a reference copy – so a copy

supplied to (eg) YouTube or PicScout allows service

providers to offer discovery or monitor pirate usage.

At present this process is ad hoc and driven by specific

business requirements (such as take-down).

Digital declaration will become the key to the

authoritative “once for all” declaration of works and

associated Rights throughout the digital network.


The future DIN: Rights IDs

Rights data needs to be shared automatically much

more than it currently is in the supply chain (volume,

multi-media, re-purposing explosion).

Policies and Rights will have IDs issued at the point at

which they are to enable resolution.

eg PLUS Coalition

Much but not all of this will be proprietary “behind

closed doors”.

Assertions (“who says”?) and Conflicts will become

critically important in managing distributed Rights data.


The future DIN: Controlled vocabularies (“Category

IDs”)

Already well established and widely used, these will be

mapped through services like the Vocabulary Mapping

Framework (VMF) and schema.org.

Apps will access Category IDs automatically and

transform them seamlessly for different domains as

required.


The future DIN: Resolvable IDs

Creation and License/Rights IDs will allow people and

systems to resolve automatically to different services –

such as content, rights metadata or licensing systems,

eg DOI/Handle already offers “multiple resolution”.

IDs will be embedded with content or in web pages, but

metadata increasingly stored remotely – especially for

rights.


“Declaration” vs “registration”

This presentation talks about “declaration” rather

than “registration” of digital content, to avoid a

common misunderstanding.

“Registration” has several meanings, and is

sometimes used to describe an act which establishes

the existence of copyright in a creation.

“Declaration” here simply means providing a

definitive, accessible copy of a creation along with

metadata for the purposes of authoritative

identification regardless of its copyright status (for

example, ”declaration” may also apply to works in

the public domain).


The future of content and rights declaration (1)

A person or organization with content and rights will

be able to upload a Creation onto a digital service

and:

- identify themselves and their roles (with IDs)

- identify their Creation(s) and links (with IDs)

- identify the rights they control and are willing to

grant (with IDs)

- this process may apply to self-publishers or major

corporations

- this process may apply to single works or complete

repertoires



The IDs they use may be standard or proprietary but

will be “shared” with the network.

Controlled vocabularies from different namespaces

will be interoperable.

Services will use apps and APIs so that for the user

the process can be more or less automatic, using a

set of default preferences.

With the RRM underpinning the standards it can be

as simple or complex (and extensible) as it needs to

be.



Similar “ID at the point of activity” processes will

support the issuing of licenses and usage reporting

and monitoring.

There are no technical barriers to this – all the

individual processes involved are commonplace and

“big data” is no longer an obstacle.

The digital explosion described at the beginning

makes this both necessary and inevitable.


Who manages the Digital Identifier Network?

Like internet, all the participants.

Threat or opportunity for IFRRO members?

New services and intermediaries will appear, some

of them growing very rapidly.

Your legacy data and systems can be an asset or a

liability.


LCC – the next step

LCC proposes to become a consortium of

standards groups (eg Editeur, DDEX, IPTC, CC,

PLUS Coalition, W3C, DOI, other ISO identifier

agencies).

Enough of these already engaged with LCC to

secure continuity.

Anyone else (eg IFRRO) may be affiliated and

participate in, propose and/or fund projects.

Project to facilitate “digital declaration at the point

of activity” will be top of the agenda.


[email protected]

www.linkedcontentcoalition.org

Thank you for your time and attention.

http://www.linkedcontentcoalition.org/

Documents

LCC and the “Digital Identifier Network”€¦ · URI2 URI7 URI1 URI20 URI4 URI8 URI21 URI14 URI19 URI16 URI17 URI18 URI9 URI10 URI11 URI12 URI13 URI15 What this looks like (ideally)