Preservation Metadata for Institutional Repositories

Preserv Project, IAM Group, School of Electronics and Computer Science,
University of Southampton, SO17 1BJ, UK
Email: sh94r@ecs.soton.ac.uk

• Digital Preservation Service Provider Models for Institutional Repositories: towards distributed services, January 2007
• Preservation Metadata for Institutional Repositories: applying PREMIS, January 2007

  Introduction

  Metadata designed for managing digital content over a long period of time is commonly referred to as 'preservation metadata', and typically informs, describes and records a range of activities concerned with preserving specific digital objects.
  
  The broad aim of digital preservation is to ensure that the content remains accessible regardless of changes in hardware and software technologies, notably: presentation formats; changes in organisational responsibilities for managing the content; and to mitigate environmental risks (e.g. Bradley 2005).
  
  What characterises many approaches to digital preservation is the implicit assumption that, from data generation to input to archiving, preservation is managed within an expert and specialised preservation environment. Digitisation projects are good examples of this approach. This paper considers the case where data creation, deposit and content management in a repository will be performed by a range of players, many non-specialists from a preservation viewpoint. Institutional repositories (IRs) are such a case, where authors of research papers, for example, 'self-archive' their works in the IR, and where the management of the IR often has no specialist preservation skills (see Hitchcock et al. 2006). It is proposed that in such cases the IR might contract preservation requirements to an external service provider. Thus preservation metadata here must inform not just long-term management of the data but also the relationship between the IR and the service provider.
  
  This paper provides a brief overview of preservation metadata, which we then seek to develop for an application involving IRs. Currently, the authoritative reference on preservation metadata is the Preservation Metadata Implementation Strategies Data Dictionary (PREMIS 2005), on which we have focussed our initial investigation. Since some seek to make IRs a malleable concept, we explain some background that establishes the role of IRs, and place IRs in a preservation context leading to the introduction of three OAIS-based models. In this case we focus on one of these models, the service provider model, to get a handle on the analysis of the PREMIS metadata set for this application.
  
  
  "It is difficult to anticipate the metadata needed to support technical and administrative processes that are not fully developed, are
  not fully tested, and in some ways, are not even fully understood. Compounding the problem is the proviso that preservation metadata recommendations must be restrained by economic realities. Creating and maintaining metadata is expensive, so any recommended
  preservation metadata elements should be backed by persuasive evidence of necessity, as well as practical means for populating them" (Lavoie and Gartner, 2005).
  
  
  In this paper we begin our investigation of supporting preservation metadata within the practical, very real and growing content of IRs, and with reference to a number of relevant business models built into our OAIS examples.
  
  

  What is preservation metadata?

  “Preservation metadata is the information necessary to carry out, document, and evaluate the processes that support the long-term retention and accessibility of digital materials." (PREMIS 2005)
  
  In terms of digital technology and the widespread creation of digital materials, preservation metadata has had a respectable period of gestation and development. Various generic approaches have been identified within projects, with the baton seemingly passing periodically from one group to another (e.g. National Library of Australia 1999; NEDLIB, Lupovici and Masanès 2000, CEDARS 2002). The  OCLC/RLG Working Group on Preservation Metadata (2002) introduced an international consensus, while implementations that emerged such as at the National Library of New Zealand (2002) were largely application-specific in nature. According to Lavoie and Gartner (2005) the earlier efforts "largely were speculative in nature, seeking to anticipate the metadata needs of programmatic digital preservation initiatives that would emerge in the future. On the other hand, development of the more recent element sets, such as OCLC, NLNZ were more closely aligned with planning and implementation of 'production' digital archiving systems."
  
  Given the changing descriptions, slightly different terminologies and the fuzziness of the overlap between preservation metadata and other forms of more widely used metadata, such as metadata for resource discovery, administrative metadata, etc., it can be quite hard to unravel the different perspectives, although a review by Day (2003a) makes a good attempt.
  
  Fortunately, a more coherent view has appeared in what is, currently, the authoritive reference on preservation metadata, the PREMIS Data Dictionary (2005). This is based, as the full name indicates, on the idea of implementation, and for the first time provides, on examination, a thorough, rigorous and comprehensive set of preservation metadata elements that a "working archive needs to support the functions of ensuring viability, renderability, understandability, authenticity, and identity in a preservation context." (Guenther 2004)
  
  

  Evolution of institutional repositories: OAIRs

  Before examining PREMIS for an IR application it is helpful to consider what IRs are and what they do, because like preservation metadata this relatively recent development has been subject to misunderstanding, confusion and attempted re-inventions.
  
  The impetus for IRs could be said to have emerged from the Open Archives Initiative (OAI) in 1999, not to be confused in preservation terms with the Open Archival Information System (OAIS; Hirtle 2001). Although institutionally-based, or more typically departmental, 'archives' were known before this, especially in areas such as computer science and economics which were served by NCSTRL and RePEc, respectively, OAI introduced the Protocol for Metadata Harvesting (OAI-PMH) to provide common services that could operate over more general, independent sites (Lynch 2001). Search is the most obvious example of such a service. OAI-PMH enables compliant sites to be interoperable, thus making institutional, rather than only disciplinary, repositories visible and viable. For the first time institutions such as universities have the ability to capture, store and disseminate copies of the published work of their own researchers. The significance of this cannot be underestimated.
  
  OAI was aimed initially at eprint archives (Van de Sompel and Lagoze 2000), and although the protocol was soon widened to handle other digital library content, the first software to support it was EPrints, on which we base our work here at Southampton. EPrints is software for building IRs that capture and provide open access to an institution's research outputs, which are deposited directly by authors in principle using the version they created, a process known as 'self-archiving'. EPrints first appeared in 2000, and an OAI-PMH 1.0-compliant version was announced on the same day this breakthrough version of the protocol was unveiled in January 2001 (Harnad 2001, OAI 2001). This application to institutional archives, or repositories, was reinforced with the emergence of DSpace software a year later. A large number of repositories have been built using software such as EPrints, DSpace and others (see the Registry of Open Access Repositories, undated).
  
  Despite the growth of such archives, interest in the wider use of repositories has lead periodically to attempts by newcomers to broaden the term 'institutional repositories' for other purposes. Rather than redefine IRs, a better starting point is the broader term, a digital repository, characterised by Heery and Anderson (2005):
  
  • content is deposited in a repository, whether by the content creator, owner or third party
  • the repository architecture manages content as well as metadata
  • the repository offers a minimum set of basic services e.g. put, get, search, access control
  • the repository must be sustainable and trusted, well-supported and well-managed
  An institutional repository, based on experience and practice, builds on the above definition and more specifically can be characterised by type of content, the source of that content, and its dissemination:
  
  • content is the research outputs of an institution
  • submission to the archive is typically by author-self-archiving
  • dissemination is primarily through common services, such as search, but in particular by OAI-PMH services
  With growing interest in digital repositories for various purposes within institutions, such archives might be more helpfully and specifically referred to as Open Access IRs, or eprints repositories (James et al. 2003). As such, where digital preservation might generally be concerned with preserving access, for OAIRs it is concerned with preserving open access.
  
  What are the implications for preservation of repositories of this type? In this context Hitchcock et al. (2005) identified two principal features:
  
  • Heterogeneous data formats (Hitchcock et al. 2006)
    
  • Low cost per item deposited. IRs must keep costs low enough not to jeopardize open access.
  This highlights the need for automation in the collection of preservation metadata, and in the efficient labelling, selection and delivery of content to preservation services.
  
  This characterisation of OAIRs for preservation can perhaps also be informed by comparison with a large-scale preservation test, the Archive Ingest and Handling Test. The AIHT practical preservation strategy will require "mechanisms for continuous transfer of content from the wider world into the hands of preserving institutions. The AIHT is designed to test the feasibility of transferring digital archives in toto from one institution to another" (Shirky 2005). This approach involving more than one agency in content management parallels our service provider model outlined below. AIHT reveals important practical experience, although there are some differences with anticipated preservation service models for IRs. For example, in AIHT:
  1. There is no scope for interaction between creator and archive
  2. There is no moderated ongoing transfer process or protocol, just a single disc of compressed data containing all files
  3. There is no business model (i.e. who is doing what for whom, and why)
  4. The scope of the test archive may or may not reflect a typical profile of an IR
  This description of what IRs are and what they represent is not to say that the role and target content of IRs won't evolve legitimately and in an informed way to serve institutional needs and research purposes, as suggested by Dempsey (2006). It is likely that other types of content, such as research data sets (Lyon et al. 2004), will be deposited and managed within OAIRs, but not all types of content produced in universities -- teaching and learning materials, administrative documents, for example -- are best stored in OAIRs. Such materials may require more specialised submission and updating facilities, and may need to restrict dissemination.
  
  

  Three OAIS preservation models for IRs

  Having defined the preservation task in terms of preservation metadata, and the target content for preservation in terms of OAIRs, we can consider the types of services that can be offered. The OAIS reference model (Figure 1a) provides a framework in which we can construct these services (OAIS 2002). At a very general level it can be seen that IRs provide a similar range of functionality as found in OAIS -- input and output, data management and storage. OAIS imposes more formality and discipline to these processes for the purpose of long-term preservation. Thus deposit becomes ingest, and we are concerned with archival storage, all enveloped by preservation planning, administrative and management roles. To understand these distinctions and these support processes, see the Cornell tutorial (2003). Information in this system is managed in packages: submission information packages (SIPs) at point of ingest, archival information packages (AIPs) in the preservation store, and dissemination information packages (DIPs) for access by users or other services.
  
  Within the types of services we could construct we wish to support a range of business models to allow IRs some flexibility in managing the preservation risk in terms of their real resources:
  
  • service provider model (service provider is OAIS, Figure 1b) the original and core project model (Hitchcock et al. 2005)
    
  • institutional model (institution is OAIS, Figure 1c), an institution may have more than one repository, e.g. EPrints-Fedora
  • software model (repository is OAIS, Figure 1d), preservation features built into IR software, e.g. DSpace

  a	b
  c	d

  
  
  Figure 1. Three preservation models based on OAIS: a, Base OAIS functional model; b, Service provider model; c, Institutional model; d. IR model
  
  
  The basis of the three service models in the formal OAIS model are apparent in Figure 1. Representations of the OAIS reference model are ubiquitous in the digital preservation literature and may differ in presentation if rarely in detail; for reference, this version (Figure 1a) was taken from a presentation by Day (2003b). The changes in the service models are shown in red and are all focussed on the ingest-data management-archival storage roles and the relations between these as shown by the connecting arrows. In the service provider model a case could be made to re-introduce the arrow connecting the service provider and the access point (e.g. EVIE 2006), depending on the agreement between the IR and service provider partners.
  
  The three models illustrated have no specific costs attached, but represent a hierarchy in terms of level of cost that might be incurred to support preservation, based on Chapman's (2003) observation: "though quantity, quality and size of the digital materials ingested has an impact on scale, the cost of long term digital sustainability correlates more to the range of digital services offered." The range of services offered in the first model is clearly potentially greater, and more flexible than the latter two, with the software model providing a baseline requirement.
  
  Other models might include the federated model (where the federation is OAIS). A prominent federated example is LOCKSS (Rosenthal et al. 2005), which focusses on journal applications rather than more heterogeneous collections such as in OAIRs. Since it would be best to pursue such a model within LOCKSS rather than re-invent it, this model is beyond the immediate concern of this project, but could be considered if the opportunity arose.
  
  

  Preservation service provider model

  The preservation service provider model was broadly outlined in terms of shared, or third-party, preservation services by Beagrie (2002), while  RLG-OCLC (2002) reported the need for third-party preservation services to fulfill the need for trusted digital repositories. This model was proposed for IRs by James et al. (2003). Referring to this as a disaggregated OAIS-compliant model, Knight (2005) extended the idea for a model-based, rather than evidence- or experience-based, analysis. Knight presented a detailed breakdown of the model and workflow from the service provider's perspective. Experience is likely to bring both more complexity and more clarity.
  
  A similar although less detailed service provider model to be adopted in the Preserv project was developed in stages in Hitchcock (2005). This model is formalised in Figure 2. A notable feature of the illustrated model is the integration of an automated file format identification tool, PRONOM-DROID, developed by The National Archives (Brown 2005). The service provider model also fits well with an OAI application, which as we have seen is core to IR software, as OAI is predicated on the data provider-service provider relationship (Lynch 2001).
  
  
  
  
  
  
  Figure 2. Schematic of Preserv service provider model, showing IR functions, format ID tool and OAI interface to preservation service provider
  
  
  As in Knight (2005), the Preserv model as presented easily lends itself to analogy with the ubiquitous OAIS representation. In terms of the main OAIS functionality -- ingest, data management, storage, dissemination, etc. -- these models highlight how responsibilities might be shared between partners. For example, in the service provider model (OAIR-SP) the IR could be OAIS-compliant, but it need not necessarily be if the service provider delivers that compliance. At the other extreme, in the software model where there is no other partner, the IR clearly has to be OAIS-aware to provide a minimal level of compliance.There are essentially three variations:
  
  • the whole illustrated model forms an OAIS unit (as in Figure 1b)
    
  • both partners -- IR and service provider -- are OAIS-compliant
  • the service provider is OAIS-compliant
  In IR terms, however, the the formalisation of the deposit interface to embrace the requirements of OAIS ingest has particular significance: "until it becomes common practice to integrate digital stewardship and preservation concerns into the entire digital content lifecycle -- especially front-end content creation -- ­ most digital preservation workflows intended to be inclusive will be reactive instead of prescriptive." (Anderson et al. 2005)
  
  In Preserv the main service provider partner is the British Library, which of course will offer an OAIS service. Thus the second of the three variations is most likely to be the case. Figure 3 shows two simplified, co-joined OAIS models representing the IR and the service provider. The OAIS administrative functions are shown shared between the two partners pending further investigation into this model to determine practical allocations. Figure 3 also explodes the service provider into a range of optional services, which are described by Hitchcock et al. (2006).
  
  
  
  Figure 3. Two OAIS repositories in Preserv preservation service provider scenario
  
  
  The model chosen enables us to analyse and select preservation metadata elements from the PREMIS Dictionary. This paper will focus on the service provider model, while the other two models will be developed in later papers.
  
  

  Developing a preservation metadata set for IRs

  "in contrast to the support for resource discovery metadata, managers of e-print repositories have practically no preservation metadata support provided by the common repository software packages." (James et al. 2003)
  
  The aim of this investigation is to rectify that omission, bearing in mind that the IR software is only one player in our preservation service provider scenario. In building on the detailed PREMIS metadata set, critically we have identified in the preliminary selection not just those elements that might be needed in an IR preservation scenario, but the sources of the necessary metadata in terms of the service provider model outlined above (Figure 2): author/IR submitter, IR software and associated tools (in this case EPrints and PRONOM-DROID), IR policy profile and service provider.
  
  There are precedents for the first three sources. Prior to producing the data dictionary, PREMIS surveyed repositories. Despite using a somewhat leading questionnaire, it discovered (Caplan 2004):
  
  • Three-quarters of all repositories obtained metadata from their depositors and the same number extracted metadata automatically by program.
    
  • Automatic extraction by repository software was most often limited to technical metadata – size, file format, and file characteristics stored in file headers.
  • Nearly two-thirds of the respondents also had some metadata supplied by repository staff, either through manual data entry or by automatic derivation from bibliographic databases.
  The validity and need for some elements will be determined by IR policies. Broadly, a preservation profile can be developed for an IR, including content formats, usage, as well as policy profiles (Hitchcock et al. 2006). Such a profile can be used to guide selection of preservation services by institution. The Preserv project has IR partners, at Southampton and Oxford, and is building profiles of both of these IRs together with profiles of other Eprints IRs.
  
  In the PREMIS data model there are five types of entity: intellectual entities, objects, events, agents and rights. While "Intellectual entities and agents are not fully described", the majority of the entries in the data dictionary involve objects and events (Guenther 2004):
  
  • Objects: "Semantic units associated with Object entities include identifiers, environment information (e.g., hardware and software), location information, technical characteristics that apply regardless of format (e.g., fixity, size, significant properties, inhibitors, creating application information), and relationships to other objects."
  • Events: "Digital provenance metadata is centered around events that have acted upon objects and is intended to record processes during the period of archival retention. Semantic units include event identifier, event type (e.g., compression, fixity check, migration, validation, etc.), event outcome, event date/time, and related agents."
  Selection and ongoing refinement of the preservation metadata set is not only concerned with managing preservation but should also consider minimising the costs of preservation actions and services. These major entity types in PREMIS, and the rights entity, can be aligned with what according to James et al. (2003) are the most significant factors affecting costs in eprint repositories:
  
  • Potential costs involved in managing proprietary formats should repositories choose to accept whatever is offered, to them (the Object elements within PREMIS)
  • The cost of creating additional metadata, particularly that associated with the technical and administrative needs for long-term management of e-prints (the Event elements)
  • The cost of negotiating rights (the Rights elements)
  Particular attention, then, should be paid to elements in these categories that will assist the efficient collection, generation and delivery of the necessary metadata to the service provider to control these costs.
  
  With regard to rights, according to PREMIS the minimum core rights information that a preservation repository must know is the permissions that have been granted to the repository itself to carry out actions related to objects within the repository. It should be noted that OAIRs could be considered a special case as far as deposit of papers published elsewhere in journal and proceedings, for example, are concerned. In these cases the IR should require only a simple licence agreement with the author. Several US institutional repositories have publicly available agreements, e.g. Caltech, California Digital Library, DSpace at MIT. For preservation purposes any rights statement should be extended minimally to allow copying and uses prescribed by the service provider. The DSpace at MIT licence includes the following clauses relating to possible preservation actions within the IR, illustrated by MacColl (2004):
  
  
  "You agree that MIT may, without changing the content, translate the submission to any medium or format for the purpose of preservation. You also agree that MIT may keep more than one copy of this submission for purposes of security, back-up and preservation."
  
  
  If an IR anticipates using an external preservation service provider as described here, the MIT example may need to be developed further in conjunction with the service provider.
  
  

  Mapping PREMIS elements to the IR-service provider model

  Within the detailed and lengthy PREMIS Data Dictionary the five entity types -- intellectual entities, objects, events, agents and rights -- are described by entries for the main elements and subelements. In this analysis we attempt to map these elements to the potential metadata sources -- IR submitter/author (via the EPrints deposit interface), IR software (within-code EPrints), file format ID tool (PRONOM-DROID), IR policy profile and service provider -- identified in our OAIR-service provider (Figure 2) model (Tables 1-5).
  
  A possible additional source of metadata is environment registries (Table 6), which are recognised in PREMIS and other preservation activities, although there are not yet any concrete examples of such registries based on the broadest, most ambitious designs that could support and source the elements identified here. PRONOM, GDFR and JHOVE are examples of more specific registry types (e.g. for file format ID and validation) that might be the basis of more expansive implementations. Representation Networks are another ongoing preservation development that may inform environment registries.
  
  The version of the mapping presented here includes the main PREMIS metadata elements without expanding on the subelements where these are assumed to follow into the same source categories unless indicated.
  
  Since PREMIS is oriented towards implementation, this requires the development of schema to define the use of certain elements within this application. According to PREMIS, the schema should, where possible, provide controlled vocabularies or codes for populating elements, rather than relying on “free text”. In addition, the schema should be adaptable to automated workflows for metadata collection and management. This analysis has not yet extended to identifying, building or including controlled vocabularies or schema that may be required by some elements within the preservation metadata set.
  
  Tables 1-6 map the PREMIS elements to the principal sources in the OAIR-SP model. It should be noted that elements are not fixed in these tables. Some may apply to more than one table, especially where related subelements are simply wrapped in a single entry, but for clarity we have not duplicated elements between tables. For example, it may be necessary for the relationship element to be informed by the IR author, but subsequent use of that information in related subelements may be the responsibility of other sources or services.
  
  Key to tables: O optional, R required, R* conditionally required, + includes related subelements from PREMIS Data Dictionary
  
  
  Table 1: From the IR submitter/author (via Eprints interface)
  

  	PREMIS metadata elements	Part of (if not main element)	PREMIS entity type	Comment	Other OAIR-SP sources
  O	creatingApplication +		Object	probably needs author (e.g. would an MS-Word file generated from OpenOffice be flagged by ID tool?)	PRONOM?
  	originalName		Object	this refers to files uploaded with the eprint submission, named by the author and recorded by EPrints as part of the upload directions; could be a DOI? Will files be renamed by SP?
  R*	dependency +	environment	Object	e.g. schema; while the submitter must indicate a relationship, the related subelements may be generated elsewhere (see also table for environment registries)
  R*	relationship +		Object	relationship between objects, high-level categorization, e.g. structural, transformation, and other types to be determined by SP	SP
  R*	linkingIntellectualEntityIdentifier +		Object	e.g. collection; while evidence of a higher entity is provided by the submitter, the ID (e.g. URI) may be generated by another service

  
  
  
  
  
  
  Table 2: From within-code EPrints

  	PREMIS metadata elements	Part of (if not main element)	PREMIS entity type	Comment	Other OAIR-SP sources
  R	objectIdentifier +		Object	identifier of the eprint record (identifiers of the digital objects are their URLs)
  	fixity +	objectCharacteristics	Object	verifies if an object has been altered. Where is fixity check first performed? Not within EPrints currently, but a script that crawls the archive comparing files with checksums is possible
  	size	objectCharacteristics	Object

  
  
  
  
  

  Table 3: From PRONOM-DROID file format ID tool
  

  	PREMIS metadata elements	Part of (if not main element)	PREMIS entity type	Comment	Other OAIR-SP sources
  	objectCategory		Object	bitstream, file, representation; this will be "implicit" in the harvesting service	IR policy
  	compositionLevel	objectCharacteristics	Object	e.g. compression, encryption, zip; EPrints won't tell you this, but a file format ID tool might	IR policy
  R	format +	objectCharacteristics	Object
  R	software +	environment	Object	software to render or use the object; SP decides which software environments are to be supported

  
  

  
  

  
  

  Table 4: From IR policy

  	PREMIS metadata elements	Part of (if not main element)	PREMIS entity type	Comment	Other OAIR-SP sources
  R	preservationLevel		Object	Depends on "preservability", cost, etc.	SP?
  	significantProperties	objectCharacteristics	Object	e.g. pdf + links	IR submitter
  O	inhibitors +	objectCharacteristics	Object	inhibit access, use or migration, e.g. encryption	IR submitter, IR policy
  R*	signatureInformation +		Object	validates submitter, for IRs e.g. identifying authors among services, authenticating material coming from a repository, etc. These appear to be fairly 'weak' needs assuming the repository and preservation services are 'secure'. The signature itself and associated elements (e.g. keyInformation) would be generated by an appropriate tool to be decided by IR/SP policy	SP policy
  R	permissionStatement +		Rights	while the author is the ultimate arbiter of which permissions to grant, the IR policy sets a framework for standardising permissions by type of object to cover preservation requirements; the SP records and formalises the management of this information (see permissionStatementIdentifier)	IR submitter, SP
  R	permissionGranted +	permissionStatement	Rights	actions the grantingAgent allows the preservation repository, using controlled values

  
  
  

  
  Table 5: From preservation service provider
  

  	PREMIS metadata elements	Part of (if not main element)	PREMIS entity type	Comment	Other OAIR-SP sources
  R	storage +		Object	direction to locate object stored in preservation repository
  R	strorageMedium		Object	e.g. tape, hard disk, CD-ROM, DVD
  R*	relatedEventIdentification +	relationship	Object	relates objects after an event, e.g. migration
  R*	linkingEventIdentifier +		Object	Use to link to events not associated with relationships, e.g. format validation, virus checking
  R	linkingPermissionStatementIdentifier +		Object	identifier for permission statement associated with the object (see permissionStatementIdentifier below)
  R*	eventIdentifier +		Event	Events are e.g. SP actions. Each event must have unique, locally-generated ID
  R*	eventType		Event	define controlled vocabulary, e.g. capture, compression, migration, decryption
  R*	eventDateTime		Event
  O	eventDetail		Event	e.g. why the event occurred
  O	eventOutcomeInformation +		Event
  R*	linkingAgentIdentifier +		Event	about an agent associated with an event
  	linkingObjectIdentifier +		Event	about an object associated with an event
  R*	agentIdentifier +		Agent	identifies the agent uniquely within the preservation repository system
  	agentName		Agent
  	agentType		Agent	from controlled vocabulary, e.g. person, organisation, software
  R	permissionStatementIdentifier +	permissionStatement	Rights	designation used within the preservation repository system
  R	linkingObject	permissionStatement	Rights	objects to which permission pertains, e.g. by IR
  	grantingAgent	permissionStatement	Rights	identifying designation for agent (IR?) granting permission, if agent is described as entity, e.g. agentIdentifier
  	grantingAgreement	permissionStatement	Rights	agreement between IR and SP, as recorded by SP	IR policy

  
  
  

  Table 6: From environment registries

  	PREMIS metadata elements	Part of (if not main element)	PREMIS entity type	Comment	Other OAIR-SP sources
  	environment		Object	omit if bit-level preservation storage
  	environmentCharacteristic	environment	Object	assessment of the described environment	IR policy, SP
  	environmentPurpose	environment	Object	uses supported by the environment, e.g. render, edit
  R*	hardware +	environment	Object	hardware components needed by software, e.g. hardware performance required, does this object require a minimum hardware level?

  
  

  Selection metadata

  At the outset of the Preserv project we had expected, given an apparently umbilical connection between preservation and selection, to find selection factors included in preservation metadata. This is not the case. PREMIS is not concerned with selection for preservation, but with content to be preserved.
  
  Selection is generally regarded to be a vital element of preservation services, for reasons of cost. In the general, hypothetical case, selection may be principled but impractical. In terms of digital content, especially Web content, new forms of content such as email lists, blogs and wikis -- and there are many others -- raise new questions about selection that simply cannot be answered from the standard reference points. In fact, the selection question presents an inverted logic as far as preservation is concerned: it is about first deciding what not to preserve in order to identify what to preserve. Simply, if used inappropriately selection could easily be counter-productive in terms of diverting greater analysis, added cost and possible mis-selections.
  
  Fortunately, IRs present a more concrete example. It could be argued that in setting up an IR an institution commits to a responsibility for all content that it admits. At least, in defining types of content that can be deposited in an IR makes the selection issue more tractable.
  
  It is also possible to identify other factors in terms of selection. Other parties may be interested in preservation of certain materials found in IRs, for example, research funders who may wish to setup alternative preservation services for these materials (e.g. PubMed UK is being set up by the Wellcome Trust).
  
  Authors may be invited  to identify content, or special features within content, for preservation. This more subjective approach has been applied to artistic and multimedia works in the PANIC project (Hunter and Choudhury 2003). Anderson et al. (2005) describe the TAG Team Questionnaire to try and build the views of creators into preservation decisions, and NLM has devised a set of permanence ratings (Byrnes 2000), informed to some extent by creators and authors, to guide selection and preservation decisions. Both examples appear more suited to management of in-house digital library materials than to IR authors and submitters, but could be adapted.
  
  In the digital environment, selection questions raise new issues that remain to be framed. So we anticipate an extension to this investigation as we attempt to combine 'selection metadata' with preservation metadata to assist the efficiency and automation of preservation workflows for IRs.
  
  

  Next steps

  "Creating and maintaining metadata is expensive, so any recommended preservation metadata elements should be backed by persuasive evidence of necessity, as well as practical means for populating them." (Lavoie and Gartner 2005)
  
  The Preserv project will work towards test implementations of the identified preservation metadata sets for each of the three OAIS-based models identified. The elements will be refined through consultation with project partners and stakeholders, and where necessary schema and vocabularies will be developed to control the data used for certain metadata elements. On the data input side we are developing institutional profiles (Hitchcock et al. 2006) to include IR policy profiles, and working directly with service providers such as the British Library and the National Archives to identify the needs of preservation services in terms of these metadata elements.
  
  This paper has focussed on mapping the PREMIS data dictionary to one of our models: the service provider model in which the service provider is an expert preservation agency external to the IR. In later papers we will explore the application of the PREMIS metadata set to the other two OAIS-based IR models - the institutional and software models - through further implementations.
  
  

  References

  Anderson, Richard, Hannah Frost, Nancy Hoebelheinrich, and Keith Johnson (2005) The AIHT at Stanford University: Automated Preservation Assessment of Heterogeneous Digital Collections, D-Lib Magazine, Vol. 11, No. 12, December
  http://www.dlib.org/dlib/december05/johnson/12johnson.html
  
  Beagrie, Neil (2002) A Continuing Access and Digital Preservation Strategy for the Joint Information Systems Committee (JISC) 2002-2005, JISC, 01 November
  http://www.jisc.ac.uk/index.cfm?name=pres_continuing
  
  Bradley, Kevin (2005) APSR Sustainability Issues Discussion Paper, Australian Partnership for Sustainable Repositories - National Library of Australia, 28 January
  http://www.apsr.edu.au/documents/APSR_Sustainability_Issues_Paper.pdf
  
  Brown, Adrian (2005) Automatic Format Identification Using PRONOM and DROID, The National Archives, Digital Preservation Technical Paper: 1, 17 September
  http://www.nationalarchives.gov.uk/aboutapps/fileformat/pdf/automatic_format_identification.pdf
  
  Byrnes, Margaret (2000) Assigning Permanence Levels to NLM's Electronic Publications, 2000 Preservation: An International Conference on the Preservation and Long Term Accessibility of Digital Materials, York, England, December 6-8
  http://www.rlg.org/en/page.php?Page_ID=244
  
  Caplan, Priscilla (2004) PREMIS - Preservation Metadata - Implementation Strategies Update 1. Implementing Preservation Repositories for Digital Materials: Current Practice and Emerging Trends in the Cultural Heritage Community, RLG DigiNews, Vol. 8, No. 5, October
  http://www.rlg.org/en/page.php?Page_ID=20462#article2
  
  Cedars (2002) Guide To Preservation Metadata, March
  http://www.leeds.ac.uk/cedars/guideto/metadata/guidetometadata.pdf
  
  Chapman, S. (2003) Counting the Costs of Digital Preservation: Is Repository Storage Affordable? Journal of Digital Information, Vol. 4 No. 2, May
  http://jodi.ecs.soton.ac.uk/Articles/v04/i02/Chapman/
  
  Cornell Tutorial (2003) The OAIS Reference Model, section 4B in Digital Preservation Management: Implementing Short-Term Strategies for Long-Term Problems, Cornell University, September
  http://www.library.cornell.edu/iris/tutorial/dpm/
  
  Day, Michael (2003a) Preservation metadata initiatives: practicality, sustainability, and interoperability, ERPANET Training Seminar on Metadata in Digital Preservation, Marburg, Germany, 3-5 September (revised)
  http://www.ukoln.ac.uk/preservation/publications/erpanet-marburg/day-paper.pdf
  

  Day, Michael (2003b) Integrating metadata schema registries with digital preservation systems to support interoperability. 2003 Dublin Core Conference, Seattle, Washington, USA, 28 September - 2 October
  http://www.ukoln.ac.uk/metadata/presentations/dc-2003/day/slides-draft.ppt
  

  Dempsey, Lorcan (2006) Networkflows, Lorcan Dempsey's weblog, January 28
  http://orweblog.oclc.org/archives/000933.html
  

  EVIE (2006) Embedding a VRE in an Institutional Environment (EVIE), Workpackage 4: VRE Preservation Requirements Analysis, to appear
  
  Guenther, Rebecca (2004) PREMIS - Preservation Metadata Implementation Strategies Update 2: Core Elements for Metadata to Support Digital Preservation, RLG DigiNews, Volume 8, Number 6, December
  http://www.rlg.org/en/page.php?Page_ID=20492#article2
  
  Harnad, Stevan (2001) Re: Eprints Open Archive Software, posting to american-scientist-open-access-forum, January 23
  http://www.ecs.soton.ac.uk/~harnad/Hypermail/Amsci/1079.html
  
  Heery, Rachel, and Sheila Anderson (2005) Digital Repositories Review, UKOLN-AHDS, 19 February
  http://www.jisc.ac.uk/uploaded_documents/digital-repositories-review-2005.pdf
  
  Hirtle, Peter (2001) OAI and OAIS: What's in a Name? D-Lib Magazine, Vol. 7 No. 4, April
  http://www.dlib.org/dlib/april01/04editorial.html
  
  Hitchcock, Steve, Tim Brody, Jessie M.N. Hey, Paul Wheatley, Adam Farquhar and Leslie Carr (2006) Building an institutional preservation profile. Preserv project paper
  
  Hitchcock, Steve, Tim Brody, Jessie M.N. Hey and Leslie Carr (2005) Preservation for Institutional Repositories: practical and invisible. Ensuring Long-term Preservation and Adding Value to Scientific and Technical data (PV 2005), Edinburgh, November 21-23
  http://www.ukoln.ac.uk/events/pv-2005/pv-2005-final-papers/033.pdf
  
  Hitchcock, Steve (2005) Capturing preservation metadata from institutional repositories. DCC Workshop on the Long-term Curation within Digital Repositories, Cambridge, July 6
  http://preserv.eprints.org/talks/hitchcock-dcccambridge060705.ppt
  
  Hunter, Jane and Sharmin Choudhury (2003) Implementing Preservation Strategies for Complex Multimedia Objects. Seventh European Conference on Research and Advanced Technology for Digital Libraries, ECDL 2003, Trondheim, Norway, August
  http://metadata.net/newmedia/Papers/ECDL2003_paper.pdf
  
  James, Hamish; Ruusalepp, Raivo; Anderson, Sheila; and Pinfield, Stephen (2003) Feasibility and Requirements Study on Preservation of E-Prints, JISC, October 29
  http://www.jisc.ac.uk/uploaded_documents/e-prints_report_final.pdf
  
  Knight, Gareth (2005) An OAIS compliant model for Disaggregated services, SHERPA-DP Report, version 1.1, 5/09/2005
  http://ahds.ac.uk/about/projects/sherpa-dp/sherpa-dp-oais-report.pdf
  
  Lavoie, Brian, and Richard Gartner (2005) Preservation metadata. DPC Technology Watch Series Report 05-01, September
  http://www.dpconline.org/docs/reports/dpctw05-01.pdf
  
  Lupovici, Catherine, Julien Masanès (2000) Metadata for long term-preservation, Nedlib Consortium, July
  http://www.kb.nl/coop/nedlib/results/D4.2/D4.2.htm
  
  Lynch, Clifford (2001) Metadata Harvesting and the Open Archives Initiative. ARL Bimonthly Report, No. 217, August http://www.arl.org/newsltr/217/mhp.html
  
  Lyon, Liz, Heery, Rachel, Duke, Monica, Coles, Simon J., Frey, Jeremy G., Hursthouse, Michael B., Carr, Leslie A. and Gutteridge, Christopher J. (2004) eBank UK: linking research data, scholarly communication and learning. In All Hands Meeting 2004, Nottingham, 31 Aug - 03 Sep 2004
  http://eprints.soton.ac.uk/8183/
  
  MacColl, John (2004) DSpace Institutional Repositories and Digital Preservation, DPC Forum on Digital Preservation in Institutional Repositories, London, 19th October, slide 5
  http://www.dpconline.org/docs/events/041019maccoll.pdf
  
  National Library of Australia (1999) Preservation Metadata for Digital Collections, 15 October 1999
  http://www.nla.gov.au/preserve/pmeta.html
  
  National Library of New Zealand (2002) Metadata Standards Framework – Preservation Metadata, November
  http://www.natlib.govt.nz/files/4initiatives_metaschema.pdf
  
  OAI (2001) Open Meeting, Washington DC, January 23
  http://www.openarchives.org/meetings/DC2001/OpenMeeting.html
  
  OAIS (2002) Reference Model for an Open Archival Information System (OAIS), Consultative Committee for Space Data Systems, CCSDS 650.0-B-1, Blue Book, Issue 1, January, adopted as ISO 14721:2003
  http://ssdoo.gsfc.nasa.gov/nost/wwwclassic/documents/pdf/CCSDS-650.0-B-1.pdf
  
  OCLC/RLG Working Group on Preservation Metadata (2002) Preservation Metadata and the OAIS Information Model: A Metadata Framework to Support the Preservation of Digital Objects, June
  http://www.oclc.org/research/projects/pmwg/pm_framework.pdf
  
  PREMIS (2005) PREservation Metadata: Implementation Strategies Working Group Data Dictionary for Preservation Metadata: Final Report of the PREMIS Working Group, May
  http://www.oclc.org/research/projects/pmwg/
  
  Registry of Open Access Repositories (ROAR), School of Electronics and Computer Science, University of Southampton (undated) http://archives.eprints.org/
  
  RLG-OCLC (2002) Trusted Digital Repositories: Attributes and Responsibilities, An RLG-OCLC Report, May
  http://www.rlg.org/longterm/repositories.pdf
  
  Rosenthal, David S. H., Thomas Lipkis, Thomas S. Robertson, and Seth Morabito (2005) Transparent Format Migration of Preserved Web Content, D-Lib Magazine, Vol. 11 No. 1, January
  http://www.dlib.org/dlib/january05/rosenthal/01rosenthal.html
  
  Shirky, Clay (2005) AIHT: Conceptual Issues from Practical Tests, D-Lib Magazine, Vol. 11, No. 12, December
  http://www.dlib.org/dlib/december05/shirky/12shirky.html
  
  Van de Sompel, Herbert, and Carl Lagoze (2000) The Santa Fe Convention of the Open Archives Initiative. D-Lib Magazine, Vol. 6 No. 2, February
  http://www.dlib.org/dlib/february00/vandesompel-oai/02vandesompel-oai.html