APPENDIX 1

NLA Guidelines for the Development and Application of a Persistent Identifier Scheme for Digital Resources

This document looks at

  • the general considerations which need to be addressed when considering persistent identifiers for online digital materials.
  • general principles guiding an NLA persistent identifier scheme
  • specific application of those principles to the various categories of NLA digital materials and digital surrogates

1. Introduction

All resources housed in a computer system have one or more identifiers. A resource will have a system identifier which may be transparent to the user, and it may have a public identifier which may be available to a user to access the resource. In some cases, the system identifier will also serve as the public identifier. Public identifiers are used on the www to access materials directly via links in other objects or databases.

Identifier systems designed for digital materials in global distributed systems such as the www generally follow a pattern consisting of two parts, a naming authority which is registered as globally unique, and a local string which must be unique to the naming authority, but may take any form. Such systems are based on a recognition of the need for maximum flexibility in the naming of resources whilst at the same time standardising to facilitate interoperability. The only way to guarantee uniqueness in a federated distributed system is to include some form of unique name or identifier for the naming authority to supply that uniqueness. (This is the pattern recommended for URNs in RFC 1737 Functional requirements for a Uniform Resource Identifier (December 1994)

The patterns for some of the currently available naming systems are:

Scheme Pattern
URI most common pattern used is <scheme>://<authority>/<path>
e.g.
http://www.nla.gov.au/somewhere/something.html ftp://www.nla.gov.au/wherever/whatever.txt
URN urn:<Namespace Identifier>:<Namespace Specific String>
e.g.
urn:issn:1035753X
urn:nbn:fi-234567
DOI <Directory code.Registrant code>/<DOI Suffix String>
e.g.
10.1045/january2000-levy 10.1001/PUBS.JAMA(278)3,JOC7055-ABST:
Handle <Naming Authority>/<Name>
e.g.
nstrl.cornell/TR93-1335
loc.pnp/12345
OAI <scheme (always oai)>:<repository id>:<local string>
e.g.
oai:nla:123454321

This pattern is based on the premise that there is no right way to identify all materials under all circumstances and for all applications. As long as the requirement that an identifier be globally unique and persistent is met, any locally defined name will work. The issue for an organisation or individual wishing to adopt a system of persistent identifiers for their resources will be what else is required of that system and what other needs of the organisation the identifier is intended to serve.

2. General Principles

2.1 General requirements

The kind of structure described above allows maximum flexibility to an organisation in the local identification of its resources, whilst allowing the identifiers to be incorporated into a global system with the addition of an appropriate naming authority component.

At this stage, only URI schemes based on the DNS (Domain Name System) can be resolved through standard web browsers.

There are a number of qualities that will be required of a persistent identifier system regardless of the type of material being described. These are:

  • Uniqueness
  • Commitment to persistence
  • Extensibility
  • Flexibility
  • Ease of use

Uniqueness

The needs of global uniqueness will be met by the naming authority component of the system adopted. However, identifiers must be unique within the context of the organisation’s digital resources and therefore globally unique when associated with the name of the organisation or authority. Many organisations use hierarchical naming systems to achieve uniqueness.

Commitment to persistence

An organisation must have a commitment to maintain the association of the current location of the resource with the persistent identifier. It is important that a resource with a persistent identifier should not be moved, or removed without updating the location information associated with the identifier.

Extensibility

The system must be extensible and capable of accommodating all resources which require an identifier.

Flexibility

An identifier system will be more effective if it is able to accommodate the special requirements of different types of material or collections. A one size fits all approach is not always the most sensible. The appropriate level of intelligence to support processes and systems for particular applications may be incorporated where it is useful and omitted in cases where it serves no purpose.

Ease of use

Although not absolutely critical, and not essential for machine generated persistent identifiers, a system will generally be more successful if it is easy to understand and apply, and if it lends itself to short and easy to use citations. It is thus advisable to avoid long and unintelligible identifiers where possible unless they are intended purely for internal use or machine processing.

2.2 Form of the persistent identifier

Within the framework of the general requirements listed above, an organisation may to use almost any form of identifier that meets its business needs and still maintain compatibility with standard naming schemes. Decisions about the form of identifier will depend on the types of materials described and the processes and system requirements associated with their use and display.

The following are some additional considerations that will need to inform the decision on the choice of the form of a persistent identifier:

2.2.1 Dumb number or intelligent identifier

A persistent identifier may be a random collection of characters with no associated semantics that contains no information about the object it identifies – a dumb number. On the other hand, it may have some intelligence, or meaning, built into it. This meaning may have varying degrees of complexity and specificity and varying degrees of risk associated with it. As a general rule, it is easier for humans to remember and use identifiers with built in mnemonics than a sequence of meaningless characters, although for machine processing this is irrelevant.

There are many reasons to build intelligence into an identifier:

  • The identifier may be easier to cite and remember e.g. DOI: 10.1045/june2000-payette is easier to remember and understand than a meaningless string of digits.
  • The identifier may assist in the checking of work from a digitisation contractor.
  • The identifier may be parsed by presentation software to display related files.
  • Intelligent identifiers are easier to make unique, especially amongst a variety of organisational units. Running numbers need to be carefully controlled and monitored across an organisation to ensure uniqueness.
  • The identifier may include relational information whereby smaller components are identified by reference to larger entities or collections of which they form a part (See Section 2.2.2)

It is unwise to use information that is not static in intelligent identifiers. Use only information that is forever associated with the item, e.g. collection ID, ISSN, hierarchical information relating to a permanent collection, date of publication, name of author etc.

2.2.2 Relational information in a persistent identifier

Hierarchy and structure of collections

One of the most common uses of intelligent identifier systems in the library world is to incorporate relational information that reflects the organisation and hierarchies of digital collections or aggregates. Smaller component such as digital page images or lower level aggregates are identified by reference to the larger collections or entities to which they belong. This form of identification is especially useful for digital surrogates of physical collections, although it can also be used to reflect the hierarchy of born digital collections such as serial issues and articles.

In the identification of digital surrogates of physical collections, relational information in an identifier can have advantages:

  • it preserves the relationship with the original collection or aggregate
  • it assists the digitisation process by facilitating the tracking of files, checking for completeness and quality control. This checking is more difficult for large collections if the files cannot easily be linked back to the original.
  • such identifiers can be interpreted by humans which makes for easier grouping and management of the files
  • the hierarchies expressed in the identifiers may be parsable by software designed to manipulate the files and generate displays for the user
  • the system provides obvious identifiers for logical aggregates which have no physical manifestation, e.g. for citation of a series, or groupings of correspondence, or the files which make up a diary
  • the collection can be navigated without software or an associated database

Relationship with the original

It is often useful to use existing identifiers associated with the original of a digitised work to derive identifiers for the digitised version. In addition to enabling unambiguous identification of the original work used to create the resource, it can facilitate the linking of finding aids and metadata inherited from the original source.

2.2.3 Support for machine processing

An intelligent identifier can facilitate processing, assist in management, or support applications such as presentation software, e.g. html pages can be generated dynamically using information encoded in the identifiers to retrieve and organise the files or, in the absence of metadata or a structural map, such information could enable all files belonging to a particular collection to be identified and listed.

2.3 Other considerations

Granularity

Decisions concerning the level of detail at which persistent identifiers will need to be assigned will depend on the perceived needs of the material. The granularity will be different for different applications and materials. For many purposes, needs can be met by citing a top level web page which serves as an entry point to a self referencing collection of web files, or by citing at the journal article, item or chapter level. However some applications such as rights management may require a finer level of detail.

All files may have a system identifier, but a simpler, easier to use identifier may be assigned as a public identifier.

Versions

Each version of an object will require a separate persistent identifier. A version may be different in a variety of ways. It may have different content, a different physical format or simply a different resolution in the same format. Whether and how to reflect the relationships between versions will be decided by the organisation. The relationship can be reflected in the identifier using some kind of version code or date or type code or in metadata.

Existing naming schemes

It may be necessary to accommodate an existing scheme. If this is the case, the existing scheme may be encapsulated in the persistent identifier.

Automatic assignment

If persistent identifiers need to be system assigned, there may be constraints that take precedence over other design considerations.

3. Naming scheme for the NLA

General principles

The NLA will apply these general principles:

  • Identifiers are designed to be globally unique
  • Identifiers will be compatible with a national system of identifiers
  • Identifiers must be unique within the context of the NLA collections and systems
  • Identifiers will be managed within the context of collection groupings, e.g. maps, web server, manuscripts, archival collection, music, both to ensure uniqueness and to facilitate processing.
  • In the case of digitised collections, codes indicating derivative types will be consistent across collections and will be derived from a standard list e.g. (thumbnail, display, master, structural map, sgml file)(See Appendix 1a)
  • Identifiers will be based on the individual requirements of specific categories of material

Collection Identifier

It is suggested that the identifier for each category of material commence with a collection identifier consisting of text nla followed by a subgroup, or collection identifier, e.g. nla.pic, nla.ms, nla.map, nla.gen, nla.mus, nla.aus, nla.arc, nla.web(?). Although this part of the identifier broadly accords with the type of material, it is primarily a designation of collection ownership and responsibility for the management of the collection. This will guarantee that the identifiers are unique both globally and within NLA collections and systems.

This identifier can be incorporated into one or more other schemes at a later date, but does not require such incorporation to make it unique.

4. Identification for specific categories of material

Broadly speaking, there are three main categories of material that the NLA makes available online and which require persistent identification:

  • Digital surrogates of hardcopy materials
  • Materials created digitally and hosted on a website
  • Materials hosted on behalf of others – digital archive

4.1 Digitised materials or digital surrogates

Each digital file created needs to be individually identified for processing and management.

The general principles for creating identifiers for digitised materials or digital surrogates are as follows. Identifiers should:

  • be globally unique even in the absence of a national resolving system.
  • incorporate aggregate and hierarchical relationship information (see section 2.2.2)
  • be based on an existing local or public identifier for the original if one exists (e.g. local identifiers such as accession number for maps or manuscripts; public identifiers such as Ferguson number, ISBN or ISSN or Kinetica number for published materials)
  • Identifies the purpose of a file (display, unstructured transcript, intermediate object, generation, etc.)

4.1.1 Collections of original, unpublished and rare materials

This category includes manuscript collections, oral history collections and some groupings of pictorial material.

Although individual items held in collections may have intrinsic value, many items derive their research value from their context within the collection to which they belong. Researchers may also cite them in the context of the collection. It is therefore important that any digitisation of collections of this kind be undertaken in such a way that the organisation and relationships that are present in the original collection can be reproduced for browsing and display purposes. (Many major digitisation projects use systems of this kind e.g. the National Digital Library Program of Library of Congress and RLG Studies in Scarlet).

For the reasons outlined in Section 2.2.2, the form of the identifier for materials within a specific grouping of material is based on an identifier for the group as a whole. If the collection already has a well-established unique identifier, it will be used as the collection identifier. This will assist in the relating of bibliographic metadata and finding aids for the original work to the digital collection.

This identifier can be extended by adding a representation of the particular aggregate(s) within that collection, if applicable, e.g. the series number and item number in a manuscript collection or sub-collection identifiers in a pictorial collection, or a series in the Oral History recordings. The identifier can be further extended by representation of further aggregates, if appropriate (item within a series)

At the lowest level, the identifier for the item can be extended by sequential numbering for the page images themselves and an extension indicating purpose of the resource, e.g. thumbnail, master copy, display copy, sgml text version.

For numbered pages, a sequence number and a page number and coded information for features such as contents lists, title pages, etc. should be added.

4.1.1.1 Manuscript collections

The naming of files in the manuscript collection will mirror the organisation of the physical collection as exemplified in the finding aid. It will reflect the logical rather than the physical collection, i.e. it will be based on the series and items within the series rather than the physical arrangement in boxes or folders.

There is a need to preserve the relationships which exist in the collections in order to allow the user of the files to navigate across, up and down the hierarchies of the series, folders and items that make up the collection structure. There is also a requirement to be able to go directly to an individual file or an individual group of files making up a digital object or sub-object that represents a particular item or grouping.

Each file will have a persistent identifier. The pattern for manuscripts identifiers will be:

<collection id>-<collection no.>-<series no.>-<item no.>-<sequence no.>-< role code>-<generation code>

<collection id>: The collection id will be nla.ms for items in the manuscript collections

<collection no.>: The top level aggregate, is the collection or accession number, e.g. MS8822 which represents the Mabo papers. This number is used as the collection location and the identifier for the finding aid.

<series no.>: The series number will be represented by a 3 digit number, with leading zeroes.

<item no.>: The item number within the series, represented by a 4 digit number with leading zeroes.

<sequence no.>: The sequence of the image within the item, represented by a 3 digit number with leading zeroes

<role code>: a code drawn from a standard list indicating the role of the file

<generation code>: A two digit code representing the version of a resource that has been migrated from its original format.

e.g. nla.ms-ms8822-001-0001-001-m
(The master file for the first page of the first item in series 1 of the Mabo papers.)

nla.ms-ms8822-001-0001-002-d
(The file for the display image of the second page of the first item in series 1 of the Mabo papers.)

In addition, an identifier may also be assigned to each logical aggregate in the hierarchy. (e.g. the collection, the series, the letter, the diary …). This aggregate may have no tangible existence as a digital object, and the use of such an identifier will depend on the application used for the presentation of the digital objects that make up the aggregate. For example it may be used to initiate the generation of a page or display of the items making up the item or it may be used to access metadata in the form of a finding aid to locate and display its component files.

e.g. nla.ms-ms8822-001-0001 refers to the first item in series one of the collection, for which there is no corresponding discrete file.

4.1.1.2 Oral History sound recordings

Sound files are generally cited and requested by the TRC number. The TRC number is an old and well-established number based on the original tape recording number. This identification number has been carried over into the naming system for the digital files as the collection is converted to digital form. TRC numbers are used as control numbers in the sound recordings database. As unique and persistent numbers already in use for digital as well as analog recordings, they are suitable for use in the online environment.

The TRC number consists of 12 digits organised into three groups of four in the pattern 0000/0000/0000. These groupings represent a hierarchy consisting of the collection number, the series within the collection and the tape number within the series.

The identifier for oral history recordings will take the form:

<collection id>-<collection no.>-<series no.>-<tape no.>-<fragment identifier>-<role code>

<collection id>: The collection id will be nla.oh for items in the oral history collections

<collection no.>: the TRC number of the recording.

<series no.>: a four digit number representing the series number within the collection, left justified with leading zeroes.

<tape no.>: a four digit number representing the tape number within the series, left justified with leading zeroes.

<fragment identifier>: formalised representation of a sound byte from the recording (format to be determined)

Note: When the details of the pilot for the delivery of sound bytes over the web are finalised, it will be necessary to examine how the fragment identifier will be further divided. Possibilities include notation indicating the starting minute and ending minute of the recording fragment within the larger recording, or the starting minute and duration of the fragment.

<role code>: a code drawn from a standard list indicating the role of the file, e.g. Transcript, master file.

<generation code>: A two digit code representing the version of a resource which has been migrated from its original format

e.g. nla.oh-trc0234-0003-0001-m
(Master recording of the first tape in series 3 of TRC0234)

4.1.2 Single items of original and rare material

This category includes maps, pictures, papers, unpublished books, and single manuscripts

If the item already has a well-established unique identifier, this will be used as the basis of the identifier. If the item does not have a suitable identifier, a running number or some other unique identifier will be used as the basis of the identifier.

4.1.2.1 Maps

The physical Rare Map collection is largely organised and named by collection or aggregate, e.g. Ferguson, Nan Kivell and Tooley. The pre 1900 Petherick and the remainder of the Library’s Rare maps are designated simply the Rare Map Collection and are assigned a running number within that collection.

The items are numbered within the collections using a unique number prefixed by a collection identifier. The collection identifiers are:

F (Ferguson)
NK (Nan Kivell)
T (Tooley) and
RM (Rare Maps)

The collection identifier is prefixed by a general collection identifier MAP, (e.g. MAPF) to further identify the material as cartographic, since some of these large formed collections also contain textual and pictorial materials.

The identifier for maps consisting of a single image will take the form:

<collection id>-<item no.>-<role code>

<collection id>: The collection id for maps is nla.map

<item no.>: The item number for maps will be the map number used in the physical collection as described above.

<role code>: a code drawn from a standard list indicating the role of the file.

e.g. nla.map-f1234-d
(a file representing a display version of the Ferguson map 1234)

The identifiers for single maps consisting of multiple segments will take the form:

<collection id>-<item no.>-<segment position id>-<role code>-<generation code>

<collection id>: The collection id for maps is nla.map

<item no.>:-The item number for maps will be the map number used in the physical collection as for single image maps.

<segment position id>: A two character code indicating the position of the item in a table representing the whole item, as follows

cp 1 horizontal co-ordinate, alphabetic beginning with a for row 1

cp 2 vertical co-ordinate, numeric beginning with 1 for column 1

e.g

nla.map-t2345-a1 nla.map-t2345-a2 nla.map-t2345-a3
nla.map-t2345-b1 nla.map-t2345-b2 nla.map-t2345-b3
nla.map-t2345-c1 nla.map-t2345-c2 nla.map-t2345-c3

Identifiers for a set of nine images that make up a digital representation of map T2345.

<role code>: drawn from a standard list

<generation code>: A two digit code representing the version of a resource which has been migrated from its original format

4.1.2.2 Pictorial materials

The identifier for pictorial images will be based on the immutable, or Kinetica record number for the bibliographic description. Identifiers will take the form:

<collection id>-<item no.>-<role code>-<generation code

<collection id>: The collection id for pictorial items is nla.pic

<item no.>: the Kinetica number of the bibliographic record for the original work. In the case of a Bibliographic record which represents a number of images in a collection, the item no. will be made unique by the addition of e suffix,

e.g nla.pic-an1234567-t. (for a thumbnail of a single image)

or:

nla.pic-an456789-1-v
nla.pic-an456789-2-v
nla.pic-an456789-3-v
etc. (for view copies of multiple items represented by a single bibliographic record)

<role code>: derived from a standard list.

<generation code>: A two digit code representing the version of a resource which has been migrated from its original format

4.1.3 Copies of published textual and music materials

4.1.3.1 Music

Identifiers for published music will follow the pattern:

<collection id>-<work no.>-<sequence no.>-<page no.>-<role code><generation code>

<collection id>: The collection id for printed music will be nla.mus.

<work no.>:- If a published work has a unique number in the public domain that can assist in the identification of the manifestation represented by the digital surrogate (e.g. a Ferguson number, Kinetica number or an ISMN), the unique number will be used as the work no. If no such number exists, a running number, or dumb identifier will be used.

<sequence no.>A three digit number with leading zeroes representing sequence of the image in the work

<page no.>- A group of four characters, the first being an alphabetic character indicating a special feature of the page, e.g. title page contents listing or index, the last three being a three digit number with leading zeroes representing the page number of the image if it has one.

If a page is unnumbered, the page no. will contain zeroes.

If there is no feature, the first character will be zero

<role code> derived from standard list

<generation code>: A two digit code representing the version of a resource which has been migrated from its original format

e.g. nla.mus-12345-002-t000-v
(A display copy of an unnumbered title page that is the second image in the group representing music item 12345)

nla.mus-45678-007-0005-v
(A display copy of page 5 which is image no. 7 in the group representing music item 45678)

4.1.3.2 Monographs

Identifiers for published monographs will follow the pattern:

<collection id>-<work no.>-<chapter or section no.>-<sequence no.>-<page no.>-<role code>-<generation code>

<collection id>: The collection id for printed monographs will be nla.aus for Australian material or nla.gen for overseas material.

<work no.>:- If a published work has a unique number in the public domain that can assist in the identification of the manifestation represented by the digital surrogate (e.g. a Ferguson number or an ISBN), the unique number will be used as the work no. If no such number exists, a running number or dumb identifier will be used.

<chapter or section no.> A three digit number with leading zeroes representing the chapter or section number

<sequence no.>A three digit number with leading zeroes representing the sequence of the image in the work

<page no.>- A group of four characters; the first three being a three digit number with leading zeroes representing page number of the image, the fourth being an alphabetic character indicating a special feature of the page, e.g. title page contents listing or index. If the page is unnumbered, the page number will contain zeroes.

<role code>: a code drawn from a standard list indicating the role of the file

<generation code>: A two digit code representing the version of a resource which has been migrated from its original format

e.g. nla.gen-0457362546-002-043-039-d
(A display copy of page 39, which is part of chapter 2 and is image no. 43 in the group representing the book)

nla.gen-0457362546-002-044-000-d
(A display copy of the following page that is an unnumbered plate which is also contained in chapter 2 and is image no. 44 in the group representing the book)

4.1.3.3 Serials

As with other digitised collections, there is, for printed serials, a need to be able to preserve the relationships which exist in the printed product in order to allow the user of the files to navigate across, up and down the hierarchies of the individual issues and volumes which make up the serial title. There is also a requirement to be able to go directly to an individual file or an individual group of files making up a digital object or sub-object that represents a particular article or issue.

The persistent identifier for a serial file will reflect the place of the file in the structure of the serial and will include date and/or issue/vol numbering and page numbering.

The pattern for serial item identifiers will be:

<collection id>-<serial identifier>-<issue no.>-<issue date>-<sequence no.>-<page no.>-<role code>-<generation code>

<collection id> The collection id for published serials will be nla.aus for Australian material or nla.gen for overseas material.

<serial identifier>If a published work has a unique number in the public domain which can assist in the identification of the manifestation which is represented by the digital surrogate, (e.g. a Ferguson number or an ISSN), use that number as the item number.

If no such number exists, obtain an ISSN and register the title with the ISDS Centre. In the case of an overseas serial, ask the ISSN Agency to obtain an ISSN from the appropriate ISDS National Centre.

<issue no.> A six digit number, the first three digits representing the volume number and the second three digits representing the numbering of the issue

<issue date.> An 8 digit number in the form YYYYMMDD representing the issue date. A cumulative index will be identified by the year of issue and the code in

<sequence no.> A three digit number with leading zeroes representing the sequence of the image within the issue

<page no.> A group of four characters; the first three being a three digit number with leading zeroes representing page number of the image, the fourth being an alphabetic character indicating a special feature of the page, e.g. title page contents listing, index. If the page is unnumbered, the page no. will contain zeroes.

<role code> A code indicting the role of the file derived from a standard list.

<generation code>: A two digit code representing the version of a resource which has been migrated from its original format

Comment:

Citation at the article level can be expressed in a number of ways:

  • Citing the identifier of the first page of the article within the issue (can be accommodated by the format outlined above
  • Assigning an article number within the issue (an addition to the format outlined above)
  • Assigning some other identifier for the article such as author’s name or initial letters of the words of the title.
  • Use of structural mapping

4.2 Materials created digitally and hosted on a website

Persistent identification of materials hosted on a website is potentially a more complex issue than that of specific groups of digitised materials. Digitised materials are likely to be a self contained, homogeneous and static group, whereas the materials on a working and evolving website will be expanding, heterogeneous and dynamic.

The challenge for persistent identification scheme for websites is to provide for flexibility and extensibility without introducing unnecessary complexity.

It is not necessary to assign persistent identifiers to all materials on a website. Materials hosted on a website will generally fall into three broad categories:

  • Transient or ephemeral materials that need neither preserving nor persistent identification. Such categories include notices of one-off events, old versions of programs of events, advertisements for materials available for sale, drafts made available for comment, etc.
  • Dynamic materials which are amended or updated frequently which are not likely to be cited but which users may bookmark to enable them to consult the most recent version. This requires only the most recent version to be accessible via a the persistent identifier for the resource. On the NLA site, categories would include What’s new, descriptions of services offered, resource listings, corporate structure of the library, contact listings, collection descriptions, navigation pages, etc. These items will Dynamic materials which are amended or updated frequently which are not likely to be cited but which users may bookmark to enable them to consult the most recent version. This requires only the most recent version to be accessible via a the persistent identifier for the resource. On the NLA site, categories would include What’s new, descriptions of services offered, resource listings, corporate structure of the library, contact listings, collection descriptions, navigation pages, etc. These items will be interesting to future researchers, but it is not necessary to meticulously save each version as they are modified. The regular snapshots of the website mandated by Australian Archives will suffice for preservation of superseded versions of this material. The library does not have at this stage dynamic files that would require sequential archiving outside this process.
  • Static and permanent materials of research value. These are the definitive versions of a resource that reflect a point in time and will not be altered. Such resources include policy papers, annual reports, staff presentations, online exhibitions, publications such as Gateways, etc. This type of material is likely to be cited, or referenced, and it should therefore be persistently identified and its accessibility maintained.

Like most early websites, the material on the NLA website is currently identified using a URL based on its place in the directory structure. Care is taken to move material as little as possible, and the latest version of the website guidelines reflects the guidelines for managing for persistence.

Nevertheless, it would be worthwhile devoting some resource to a systematic analysis and classification of the information on the site and an identification of the files that need to be retained indefinitely and allocated a persistent identifier.

There is no need to take any action at all for the transient information on the web site. Its transitory nature means that once its useful life has gone, it can disappear.

Dynamic information, as described above needs to be accessible for long periods of time at the same location, but not indefinitely. It does not need to have a life beyond the current organisational and technical environment and therefore does not need a location and protocol independent identifier system. It is therefore a good candidate for well managed URL style URIs.

Static and permanent information of research value on the other hand needs to be identified and made accessible indefinitely. A system of persistent identifiers is therefore appropriate to this material. Ideally an identifier should be assigned when the material is loaded to the server so that citations to the material use the identifier as soon as the material is made public.

Much web material is self-referencing and therefore does not need grouping. Nevertheless, it might help to manage the files at a later date if files that are part of the same digital object are grouped or related by a persistent identifier.

Persistent identifiers for the archival component of the website should be assigned at the time they are made available online, according to the principles outlined in Managing web resources for persistence. They should be candidates for use of the identifier to be developed by the NLA in conjunction with the state libraries.

4.3 Persistent identification of digital archival material (Pandora)

Material destined for a digital archive requires persistent identification. By definition, it is valuable material, preserved for posterity and therefore needs to be accessible indefinitely via an actionable identifier.

NLA’s current digital archive was developed as an outcome of the Pandora project and consists largely of material harvested from selected websites. Many of the sites are harvested periodically and when this is done, all files are retained even though the content of many files remains the same.

A large number of files are harvested and the volume of files processed makes it impractical to manually assign identifiers to each component file. It is therefore necessary to automatically assign identifiers to each file.

Identifiers for materials in the NLA digital archive consist of a work identifier, the date of archiving the particular instance of the resource, and the host, path and file name of the original file. A combination of the work identifier and the date harvested relate the individual files that make up the particular version of the resource as it existed at that point in time. The work identifier relates the different versions harvested on different occasions and will be used as the identifier for the Pandora title entry page.

Specifically, the Pandora PIs take the form:

<collection id>-<work identifier>-<archive date>-<publisher’s URI>-<generation code>

<collection id>: For the archival collections the collection ID will be nla.arc

<work identifier>: a unique number within the digital archival collections assigned to the parent work of which the resource is a component.

<archive date> the date the file was archived in the format YYYYMMDD.

<publisher’s URI> currently the host name, path name and file name of the resource on the publisher’s site.

<generation code>: A two digit code representing the version of a resource which has been migrated from its original format

The identifier for the Pandora title entry page will take the form:

<collection id>-<work identifier>

The above pattern results in a persistent unique identifier for each file which has the following properties:

  • It provides a unique identifier which meets the essential need for managing files
  • It provides the granularity of identification required for facilitating preservation activities
  • It contains sufficient intelligence to enable a measure of grouping and relating of versions without the availability of metadata for structural mapping

However, there remains the problem of specifying and recording the relationship between the original and the archival copy and giving practical expression to that relationship in the ability to translate citations made to the original work into requests for the archival copy.

The main problem facing an archive of resources that have not originated with the archival institution relates to consistent citation of the works archived.

Assuming that the publisher’s site is the preferred site while it exists, it follows that the identifier used on the publisher’s site will be the one used in references to the resource in other works, etc. Provided the publisher’s site is extant, this presents no problem. However it becomes vital that the identifiers for the publisher’s version and the archival version are linked, or are the same, if the full purpose of the archiving is to be achieved, i.e. that the resource remain available and recognisable in citations. There are a number of options for achieving this

  • Through metadata
  • Through multiple resolution capabilities
  • Through redirection
  • Through contractual arrangements or agreements between publisher and custodial agency

The particular method used to resolve this problem is not critical, however it is critical that the issue be tackled.

APPENDIX 1a
Role codes preliminary list

This is a standardised list intended for use across media types. New roles will be added as necessary

Code Description
t thumbnail
v view
e examination
m master
a archive
s summary
st structured transcript
ut unstructured transcript
sm structural map
fa finding aid

APPENDIX 1b
Code list for features recorded for Music items

Highlighted codes will probably be most used for music items

C Front cover
T Title page
L List title page
B Back Cover
U Contents page
I Index
X Additional page/s inserted/pasted into item
A Additional title page
F Inside (ie verso) front cover
E End papers
M Fly leaf
G Inside (ie verso) Back Cover
K Preface
R Introduction
S Port/s
Q Leaf / leaves of Plate/s
W Verso plate/s (reverse side of plate)
H Illustration/s
D Bibliography
N Biography
J Discography
V Various covers (ie. the cover of the 2nd or extra copy of the same bibliographic item that is different to the cover of the item digitised. eg. different performer featured)
1 page 1
2 etc. page 2 etc.
i page i
ii etc. page ii etc.



Music code list in alphabetical order

A Additional title page
B Back Cover
C Front cover
D Bibliography
E End papers
F Inside (ie verso) front cover
G Inside (ie verso) Back Cover
H Illustration/s
I Index
J Discography
K Preface
L List title page
M Fly leaf
N Biography
Q Leaf / leaves of Plate/s
R Introduction
S Port/s
T Title page
U Contents page
V Various covers (ie. the cover of the 2nd or extra copy of the same bibliographic item that is different to the cover of the item digitised. eg. different performer featured)
W Verso plate/s (reverse side of plate)
X Additional page/s inserted/pasted into item
1 page 1
2 etc. page 2 etc.
i page i
ii etc. page ii etc.

NOTE: Advertisements are not being coded

Return to Table of Contents | Proceed to Appendix 2

Back to the top