Preparing for Digital Preservation What is being preserved

Preparing for Digital Preservation What is being preserved: Identification and Rights Management issues Norman Paskin International DOI Foundation doi>

Recommended background material doi> • Preservation Management of Digital Materials – The Handbook – – – N. Beagrie/M. Jones/DPC www. dpconline. org/graphics/handbook/ 3. 4 Rights management 4. 4 Metadata and Documentation 4. 5 Access • Digital preservation: an introduction to the standards issues surrounding the deposit of non-print publications – M Bide/E J Potter/A Watkinson – www. bic. org. uk/digpres. doc Sept 1999

Outline of presentation doi> 1. Identifiers – 1. 1 Identifiers and metadata – 1. 2 Interoperability – 1. 3 Different meanings of “identifier” – 1. 4 Persistence 2. 3. “Keep a copy” - ? Rights – 3. 1 Accessing “definitive copy” – 3. 2 Rights framework

1. 1 Identifiers and metadata doi> • An identifier = an unambiguous string denoting an entity 0550 10234 5

1. 1 Identifiers and metadata doi> • An item of metadata = “a relationship that someone claims to exist between two entities” (indecs), each of which may have an identifier: 0550 10234 5 Pantone 4567 [Book. Data says] the cover of this book is red

1. 1 Identifiers and metadata doi> • To be useful, an identifier requires some metadata: 0550 10234 5 [Books in print says] Chambers Dictionary The title of this identified book is….

1. 1 Identifiers and metadata doi> • entity: something that is identified – “Nothing exists until is identified” – Entities may include: • Abstractions (red); technical means (MP 3 player); labels (title); things (book) etc. • ontology: structured relationships between entities – “an explicit formal specification of how to represent the entities that are assumed to exist in some area of interest and the relationships that hold among them” – (such as: “page” is component of “book”) – Examples: indecs framework; ONIX; FRBR

1. 2 Interoperability doi> • In a distributed environment, there is no one central physical archive • A distributed virtual archive requires that all the players and components interoperate

1. 2 Interoperability doi> • Across media – books, serials, audiovisual, software, abstract works, visual material, etc • Across functions – cataloguing, discovery, workflow, rights management, archiving • Across levels of metadata – Simple, complex • Across linguistic and semantic barriers • Across territorial barriers • Across technology platforms

1. 2 Interoperability doi> • Preservation: "How do we interoperate with the future? “ • Preservation issues (identifiers, metadata, rights) are the same as any other interoperability problem

1. 3 Meanings of “identifier” doi> • [1] Labels: the output of “numbering schemes” – ISBN: ISO 2108: 1992 International Standard Book Numbering – ISSN: ISO 3297: 1998 International Standard Serial Number – ISRC: ISO 3901: 2001 International Standard Recording Code – ISRN: ISO 10444: 1997 International Standard Technical Report Number – ISMN: ISO 10957: 1993 International Standard Music Number – ISWC: ISO 15707: 2001 International Standard Musical Work Code – ISAN: Draft ISO 15706 International Standard Audiovisual Number – V-ISAN: Draft ISO 20925 Version Identifier for audiovisual works – ISTC: Draft ISO 21047 International Standard Text Code – PII: Publisher Item Identifier – etc

1. 3 Meanings of “identifier” doi> • [2] “infrastructure specifications”: specifying how to make labels actionable • Do not generate a label, but if you have one, specify how to use it in some particular context – URN: Uniform Resource name – URI: Uniform [Universal] Resource Identifier – PURL: Persistent Uniform Resource Locator e. g. ISBN as URN Note same concept in also in other non-digital contexts e. g. ISBN as EAN (978…. ) bar code or RFID

1. 3 Meanings of “identifier” doi> • [3] “implemented systems” • Implement labels, through actionable specification, in a managed way – EAN/UPC: physical product codes : – implement ISO bar codes, RFIDs in the supply chain – DOI: digital object identifiers : – implement URN/URIs in intellectual property (+metadata, policy) doi>

1. 3 Meanings of “identifier” doi> “For use on the Internet, an ISBN label can become a URN specification; an ISBN label can be incorporated into a DOI, which is an implemented identifier system following the URI specification. ” Is clearer than “an ISBN identifier can become a URN identifier; an ISBN identifier can be incorporated into a Digital Object identifier, which is an implemented URI identifier” (? ) A particular use of the word may be a mix of meanings [1], [2]& [3]

1. 4 Persistence doi>

doi> URL Content URL URL URL Printed identifiers, bookmarks, etc

doi> URL URL URL 404 File not found URL URL Content URL URL "Linkrot": recent estimates 16% in 6 months

URL DOI DOI URL DOI Redirection (resolution) e. g. DOI doi> Assigner Content DOI URL DOI directory DOI URL DOI Content DOI URL URLDOI URL

doi> DOI Assigner DOI Response Page DOI DOI directory DOI DOI DOI • purchase content • view free excerpt • get related items • get archive copy • request permissions Content DOI More than just "locate" DOI DOI

doi> DOI Assigner DOI Response Page DOI DOI directory DOI DOI • purchase content • view free excerpt • get related items • get archive copy • request permissions DOI DOI DOI Archive

1. 4 Persistence doi> • Persistent identifier – Resolution (redirection) • Persistence of the associated metadata • Persistence of the resolution system • Persistence of the identified copy – digital preservation: migration, emulation, encapsulation • Persistence is a matter of social infrastructure – Technology can help but not guarantee

Internet: DOI, URN, URL, PURL doi> • Distinguish two issues: 1. The technical specification of “what is” a URN and a URI etc. – identifiers in sense [2] 2. What this means for practical implementation – identifiers in sense [3]

Internet persistent id specs doi> • See DOI Handbook – 4. 9 DOI as a URI – 4. 10 DOI as a URN – 6. 10 DOI and PURL • Aim: persistent across time and unique across network space; useful and implemented • PURLs are tied to http and are single redirect etc. • URI/URNs are intended to be abstract names – independent of protocols (approx) – DOIs are URIs (formal specification) – DOIs are URNs (in effect) • URN and URI proponents disagree • (& there are other proposed specs e. g. ARK)

Internet persistent id specs doi> http: // www. w 3. org/addressing (But largely from IETF, W 3 C did not see need for URN) URI URN URL ftp: gopher: http: urn: Resolution (N 2 L)

DOI as URI doi> • IETF formal spec “URI scheme for Digital Object identifier” – Paskin, Norman; Neylon, Eamonn; Hammond, Tony; Sun, Sam; Uniform Resource Identifier (URI) scheme for Digital Object Identifiers (DOIs); An abstract specification (uri: doi: ) – Would be doi: (like tel: ) [uri: is not part of the uri spec, unlike urn: ] • May be a pure name or de-referenced by any service – The namespace provides its own mechanism (“Bootstrapping”) • On its own, it’s just a specification! – Requires code distribution for any implementation

DOI as URN • URN is less clear: doi> – Higher level situation muddy – Set of IETF drafts that define URN – Set of registered namespaces (e. g. isbn) • DOI could be but isn’t- no advantage • Unlike URI, provides a specific DNS-based middle layer (RDS) to find the appropriate resolution service • Scalability and security questioned; and: • Little or no resolution implementation – urn: isbn: 123456789 can be defined ; but what does it do over and above isbn: 123456789? – neither have a readily available, well known, global, resolution • A DOI is more than URN or URI – Adds Policy, business rules, business model – Adds Metadata specifications (cf ISBN, EAN, Visa)

Outline of presentation doi> 1. Identifiers – 1. 1 Identifiers and metadata – 1. 2 Interoperability – 1. 3 Different meanings of “identifier” – 1. 4 Persistence 2. 3. “Keep a copy” - ? Rights – 3. 1 Accessing “definitive copy” – 3. 2 Rights framework

2. “Keep a copy” doi> • Digital preservation is “keeping a copy” – – What is it you are archiving? (or managing, or counting) What’s a copy? Something that is “the same as ” Is A the same as B? Consider a photocopy…. text; author; work; paper; spatial location…. …. etc A B

2. “Keep a copy” doi> • • “Is A the same as B? ” is meaningless Can only say “Is A the same as B for the purpose of…? ” “the same” for some is “two different things” for others Purpose is defined by attributes – “Nothing exists until is identified” – …and its relevant attributes identified • Structured metadata is needed (e. g. ONIX for digital preservation? ) A B

3. 1 Accessing the definitive copy doi> • “How can an identifier be used to locate a specific local copy, which may have different access rights? ” [see www. doi. org, FAQ 26] • Resolution of identifiers to global services. • Contextualization of requests to those services to local requirements. – split this into separate global and subsequent delegated local resolution steps e. g. Open. URL – a globally-maintained database is clearly the wrong place to hold information on every local collection. – ("Linking to the Appropriate Copy: Report of a DOIBased Prototype"; (O. Beit-Arie, et. al. ) D-Lib Magazine, www. dlib. org September 2001) • A definitive archive copy could be separately identified (with its own DOI) – a matter of policy – Functional granularity

3. 2 Rights framework doi> • ISO/IEC MPEG-21 as exemplar • Digital item: a structured digital object with a standard representation, identification and metadata • The fundamental unit of distribution & transaction in the MPEG-21 framework • Maps to “Digital Object” (DOI, Digital Object Architecture) or “Resource” (IETF) • "Digital objects provide a means of organizing and identifying content for purposes of storage, access or distribution… …metadata may include restrictions on access to digital objects, notices of ownership, and licensing agreements…" (www. xiwt. org/documents/Manag. Access. html)

3. 2 Rights framework Use DRM Technology Platform Application layer Enforcement of rights & permissions Rendering, environment etc. Expression layer Rights Expression Language Machine-capable interpretation of rights: XRML etc Vocabulary layer Rights metadata Dictionary Metadata set 1 Metadata set 2

3. 2 Rights framework doi> • Standards infrastructure must accommodate many different components (MPEG 21 standard is many parts) • But a structured digital object with a standard representation, identification and metadata is "The fundamental unit” • Must be interoperable with existing metadata standards - e. g. ONIX, SMPTE so need Dictionaries • MPEG 21 Rights Data Dictionary & Rights Expression Language – Purpose: "To achieve the goal of expressing rights for all Users of MPEG-21’s Digital Items"

Describing rights using (meta)data doi> Primary rights events (claims, deals) are described using pieces of data: Rights Statement (“claim”) [party] owns [right] in [creation] in [time] and [place] Rights Agreement (“deal”) [party] agreed with [party] in [time] and [place] that [event] Pieces of "rights metadata" used in each semantic structure

Describing rights using (meta)data doi> Primary rights events (claims, deals) are described using pieces of data: Rights Statement (“claim”) [party] owns [right] in [creation] in [time] and [place] Rights Agreement (“deal”) [party] agreed with [party] in [time] and [place] that [event] Creations typically have standard identifiers, which may have associated structured data, or which may act as keys to get this data Other pieces of data also need standard identifiers (time, party. . )

What is "rights metadata"? A mix of data from many sources: Rights “events” doi> Statements, agreements, transfers, permissions, prohibitions, requirements, assertions, approvals

What is “rights metadata”? A mix of data from many sources: doi> Rights “events” Descriptive metadata Creations, Creation types, contributor roles, user roles, tools, classifications, measures

What is “rights metadata”? A mix of data from many sources: doi> Rights “events” Descriptive metadata Legal metadata Rights, persons, intellectual property

What is “rights metadata”? doi> A mix of data from many sources: Rights “events” Descriptive metadata Legal metadata Financial metadata Terms, conventions These sets of “rights metadata" are standardized and maintained in different places.

Distributed rights management doi> This mix of data from many sources is used in many different places by different people in chains of rights events: statement assertion agreement permission transfer agreement permission prohibition requirement etc [party] can [verb] [amount] to [creation] at [time] in [place]. Each entity can be expanded to reveal more data

Distributed rights management statement assertion agreement permission doi> transfer agreement permission prohibition requirement etc Each of these is an information object –an entity - which may need to link to or use information objects in other databases. The information used by each must therefore be standardised/interoperable

3. 2 Rights framework doi> Is there a way of getting to this "interoperation of data from many sources"? Yes: work already done which shows how

indecs (www. indecs. org) doi> • Interoperability of Data in E-Commerce Systems • Produced principles for structured metadata and basis for a data dictionary for interoperability • Principles used by DOI, ONIX, etc • Applicable to other structured approaches e. g. SMPTE (and creates means of interoperability with them) • Now extended to rights transactions: – <indecs>2 rdd Consortium (includes IDF) – Accepted as basis of MPEG-21 Rights Data Dictionary

The MPEG-21 RDD doi> • A data dictionary is a place where the process of semantics definitions meets technology – MPEG standards have traditionally been about engineering solutions • MPEG-21 is a multimedia and a lifecycle framework: its rights terminology does not exist in a vacuum – Interacts with a large number of existing and developing schemes and systems – The number of terms involved is likely to grow steadily and significantly • MPEG-21 is taking the lead in establishing an RDD; it is likely to be widely supported if it is flexible and interoperable

Rights & description are interdependent (1) doi> • “Rights” metadata describes what people can (or can’t) do with assets, and when, where, how and with what they can do it. • “Descriptive” metadata describes what people did with assets: the same thing, but in the past. The majority of terms are common. • Any descriptive term may be relevant to the conditions of an agreement • When new works are created through derivation, aggregation or copying, new descriptions are needed which rely on both descriptive & rights metadata

Rights & description are interdependent (2) doi> • Ownership changes and changes of law or jurisdiction often require querying of descriptive metadata for implementation in systems • “Requirements” can be dependent on description in complex (and unfamiliar) ways • Terms from descriptive schemes such as ONIX, Mi 3 P, DOI-NS, PRISM, MPEG 7 Descriptor Schemes, DC and SCORM (and many others) will need to be integrated with any effective RDD

Relationship with other metadata schemes doi> • Many content metadata schemes are in use and development and there will be many more • These all impact on rights descriptions. Users will be reluctant (or unable) to adopt separate terms for “rights” descriptions – automated interoperability into and out of RDD terms needed • Users need to describe “non-digital” rights in tandem with digital • The meaning of terms in external schemes must be fully mapped to RDD terms so that they form a part of the available data dictionary and enable users to automate their participation

<indecs> Data Dictionary doi> • To provide a method for generating a set of clear, consistent, structured and integrated terms and definitions, to the required level of granularity, for an MPEG Rights Data Dictionary • To provide a comprehensive methodology for the interoperability of terms from different schemes and systems used in the management of rights and permissions through mapping. – Will be used by DOI Application Profiles – DOIs can deliver this required interoperability • To describe but in no way prescribe how rights and permissions operate • To provide a framework for future governance.

Outline of presentation doi> 1. Identifiers – 1. 1 Identifiers and metadata – 1. 2 Interoperability – 1. 3 Different meanings of “identifier” – 1. 4 Persistence 2. 3. “Keep a copy” - ? Rights – 3. 1 Accessing “definitive copy” – 3. 2 Rights framework

Additional material doi> • “DRM Technology: Identification and Metadata” – Norman Paskin – In: Digital Rights Management: Technical, Economic, Juridical and Political Aspects (ed. Becker et al) – Springer Lecture Notes in Computer Science series – In press • "Towards a Rights Data Dictionary - Identifiers and Semantics at work on the net". – Norman Paskin – imi insights, June 2002 – http: //www. epsltd. com/IMI. htm (subscription access) • Copies available from author on request (n. paskin@doi. org)

Norman Paskin, International DOI Foundation n. paskin@doi. org doi>