Content packaging and MPEG-21 DID Andy Powell UKOLN

Content packaging and MPEG-21 DID Andy Powell, UKOLN, University of Bath a. powell@ukoln. ac. uk JISC Joint Programmes Meeting, July 2005, Cambridge UKOLN is supported by: www. bath. ac. uk www. ukoln. ac. uk a centre of expertise in digital information management JISC Joint Programmes Meeting

Contents • why • how PS - I’m only going to tell you how using MPEG-21 DID. PPS - I’m not really going to tell you how to use MPEG-21 DID since until 24 hours ago I knew pretty much nothing about it, but many thanks to Jeroen Bekaert and Herbert Van de Sompel of LANL for sharing their presentations and thoughts with me… JISC Joint Programmes Meeting

Why… • … do we need packaging standards? • because nothing in life is simple! • applications tend to be built around aggregate objects of one kind or another – digitised books (multiple parts, i. e. pages) – learning objects (pathway thru multiple objects) • even objects that look simple (e. g. an eprint in an eprint archive) turn out to be complex • we need to exchange ‘complex objects’ JISC Joint Programmes Meeting

Packaging standards http: //dev. dcc. rl. ac. uk/twiki/bin/view/Main/Content. Packaging • framework for encoding ‘complex objects’ – containing multiple objects and/or metadata – may be nested • often using XML as the encoding framework • component objects embedded ‘by value’ – XML fragments or Base 64 encoding or … • or passed ‘by reference’ – URI • like all standards, there are plenty to choose XFDU = XML Formatted Data Unit CCSDS = Consultative Committee for Space Data Systems from: – METS, IMS-CP, MPEG-21 DID, XFDU (CCSDS) JISC Joint Programmes Meeting

Example 1 – digitised book • consider digitised book in a ‘digital library’ • complex object (book) made up of component parts (pages) • the object and the parts each have one or more chunks of associated metadata • parts may be embedded (by value) or linked (by reference) book • typical application: ‘virtual page turner’ page • typical packaging format: METS JISC Joint Programmes Meeting

Example 2 – learning object • consider a learning object (whatever that is!) • complex object made up of one or more parts (other learning objects) – documents, images, videos, interactive objects, audio files, maps, etc. – each with IEEE LOM metadata – some notion of workflow or sequencing ‘thru’ the parts learning object • typical application: VLE • typical packaging format: IMS-CP JISC Joint Programmes Meeting video PDF document multiplechoice test

Example 3 - eprint • eprints in eprint archives typically exposed using OAI-PMH currently • repositories take different approaches to what they expose in their metadata – work, or manifestations, or both eprint (work) PDF (manifestation) manifestation MS-Word (manifestation) now: expose separate simple objects (metadata only) PDF Word future: expose complex objects (metadata and full text) JISC Joint Programmes Meeting

Some observations… • complex objects made up of component parts and metadata – component parts need not be digital objects • common approach to modelling objects crucial to interoperability • unfortunately… a common approach to modelling is difficult to achieve – especially on an international scale • but no ‘best practice’ yet for how to model the complex objects found in eprint archives and institutional repositories JISC Joint Programmes Meeting

MPEG-21 • ISO standard that defines ‘a normative open framework for multimedia delivery and consumption for use by all the players in the delivery and consumption chain’ • MPEG-21 is modular (17 parts!) – – – Part 2: DID – representation of digital objects Part 3: DII – identification of digital objects Part 4: IPMP – enforcement of rights expressions Part 5: REL – declaration of rights expressions Part 7: DIA – transcoding based on contextual information – Part 10: DIP – association of behaviors – Part 16: BF – binary representation of digital objects • parts can be (and are) used autonomously JISC Joint Programmes Meeting

MPEG-21 DID • ISO/IEC 21000 -2: Digital Item Declaration – second edition ISO standard • representation of Digital Objects – referred to as Digital Items • 3 sections – abstract model – DID – representation of the Model in XML (DIDL) – W 3 C XML Schema • other representation of DID may emerge – ISO/IEC 21000 -16: Binary Format – Resource Description Framework (RDF)? JISC Joint Programmes Meeting

DIDL and digital libraries • development of MPEG-21 largely driven by the big movie/video content owners and broadcast industries • little widespread involvement by digital library community • however, significant involvement of staff at LANL Research Library – input into 2 nd edition of the standard – ISO to publish the standard for free on the ISO Web site JISC Joint Programmes Meeting

MPEG-21 DID model and/or identifier instantiated as DIDL/XML document metadata container Descriptor/Statement item Descriptor/Statement (sub-)item Descriptor/Statement component Descriptor/Statement resource (datastream) items (with optional nested sub-items) made up of one or more components, each of which has one or more datastreams represents the digital item/object/asset JISC Joint Programmes Meeting 12

Example – journal article • consider a ‘journal article’ digital asset comprising a PDF document and an XML metadata record (both by reference) • the asset is mapped to a DIDL ‘Item’ element as follows… DIDL elements Item Descriptor/Statement Component Descriptor/Statement Resource @mime. Type @ref secondary info type value dii: identifier doi: 10. 1234/5678 dcterms: created 2005 -02 -22 2005 -02 -2 application/xml http: //example. org/get. Meta? id=5678 dcterms: created 2005 -02 -21 application/pdf http: //example. org/get. PDF? id=5678 JISC Joint Programmes Meeting 13

MPEG-21 DIDL strengths • MPEG-21 DID ‘abstract model’ provides a syntax-independent way of – defining underlying ‘complex object’ model in use – mapping between complex objects from different repositories • combination of DID and MPEG-21 Digital Item Identification (DII) allows for unambiguous global identification of complex objects and their component parts JISC Joint Programmes Meeting

MPEG-21 DIDL strengths (2) • DID model (and associated DIDL XML schemas) provide open, extensible framework • in particular, the Descriptor/Statement mechanism provides a way of attaching arbitrary metadata (secondary information) to the digital object or any of its component parts – – checksums and/or signatures DC description rights metadata (MPEG-21 REL) etc. JISC Joint Programmes Meeting

Final thoughts… • some danger that discussion about packaging formats is a religious issue (i. e. don’t bring it up in a job interview!) • the really difficult ‘issues’ are in the way that our complex objects are modelled – applies to all the packaging formats, i. e the problem is independent of chosen packaging standard • need widespread agreements about how we model our complex objects in order to achieve interoperability • need to do that on an international basis JISC Joint Programmes Meeting

Questions? JISC Joint Programmes Meeting 17