In an article written in Finnish, it is better to use phrases “Julkaisussa” (“In” in Finnish) and “Saatavissa” (“Available from” in Finnish) instead of “In” and “Available from”.

For a publication written in Latin script, any data element in a non-Roman alphabet in the cited information resource is Romanized, with the Romanized portion appended to the data element as associated information.

For a publication not written in Latin script, any data element in a Roman alphabet in the cited information resource is transliterated or transcribed to the script used, with the converted portion appended to the data element as associated information.

==== Roles

• Serialisation: bibitem/contributor, Person, Organization, Affiliation

==== Selection

==== Personal names

===== General

===== Variant forms

===== Additions

==== Organization or group names

===== Form of name

===== Disambiguation of names

===== Subordinate body

===== Parent body

==== Multiple creators

===== Title first

===== Conferences

===== Subsidiary creator

===== Originators of multiple editions

===== Derivative work creator

==== Pseudonyms

==== Anonymous works

4.1 classifies the various possible roles a creator can have with regard to a resource into six broad categories in order of priority, with more refined classifications typically named for the creator. These classifications are retained in the model: bibitem/contributor/role@type provides the broad category of role (author, performer, compiler, adaptor, publisher, distributor), while bibitem/contributor/role/description contains one or more specific descriptions of the role of the contributor.

The content of the bibitem/contributor/role element is optional, and contains a more specific expression of the creator role. The description can be strictly a noun, e.g. “librettist”, or it can be a phrase, corresponding to how it will be rendered in a citation (“libretto by”); the choice is at the discretion of the bibliographer, and is determined by how the citation is expected to be rendered (as rendering systems are unlikely to be sophisticated enough to convert the one into the other.)

Whether subsidiary creators should be included in a bibliographic item, and which creators should be included, is at the discretion of the bibliographer.

Exceptionally, “translator” is differentiated from the other “adaptor” roles, because it is so different from other adaptor roles such as engraver or photographer.

For derivative works, the creator and title of the original resource can be mentioned as a bibliographic note (bibitem/note); however best practice is to treat the original resource as a distinct related bibliographic item, related through the “derivedFrom” relation (). Likewise, different originators of different editions should be modelled by modelling the editions as distinct related items, each with its own distinct editor as contributor. (For example, the later edition given as the main item, and the earlier edition related through a “updatedBy” relation.)

Although both composer and librettist are included as authors under 4.1, the two roles are clearly differentiated in 5.7, with the composer given priority over the librettist.

=== Personal names

Names may be modelled either broken up into their constituent components (prefix, forename, initials, surname, addition), or as a complete name string. The latter is useful if (contrary to the recommendation of 4.1) language-specific conventions for ordering of name components are to be observed; e.g. “James CLARK, John COWAN, MURATA Makoto”.

If there are multiple contributors named in a role, all contributors are expected to be represented in the bibliographic entry. This reflects practice in electronic bibliographic tools, and renderers may choose to truncate contributors with “et al.” beyond a certain cuttoff point.. If this information is not available, a dummy contributor can be named with “et al.” as their completename, but this is not preferred.

The initials component is used if all forenames are replaced by initials. If only the middle name is replaced by an initial, it is still treated as a forename.

=== Organisations

The model caters for both full names and abbreviations of organisations. Subordinate body names are modelled separately from the parent body name; the choice of whether to model parent + subordinate or a single name depends on the degree of independence of the subordinate body, as outlined in 4.1.

=== Pseudonyms and Variant Names

Variant names, and assumed names, can both be modelled as notes on names. However, for greater clarity, and for modelling of the internal structure of the variant or assumed name, a variant name can be modelled, with the same level of detail as default names. The type of variant, or relation between the name and variant, needs to be made explicit in that case. For example pseudonym/pseudonymOf for literary pseudonyms, or username/usernameOf for social media usernames.

Variant names can be applied to personal names, organization names, and organization subdivision names.

=== Anonymous and Varii

Descriptors of authors, such as “Anon” or “Various Authors”, should be given as completename elements. The model does not standardise the designations of authors. As with pseudonyms, the real name of the creator if known can be modelled as a note, or as a name variant.

== Contribution and contributor models

Contributor model

Person/name

The name of a person.

Person/name/forename

A forename of the person.

Person/name/initials

The initials of the person (initials of forenames and middle names).

Person/name/surname

A surname of the person.

Person/name/addition

A suffixed addition to the name of the person.

Person/name/prefix

A prefixed addition to the name of the person.

Person/name/completeName

The preformatted version of the name of the person, not broken down into its component parts.

Person/name/note

An additional note about the name of the person

Person/name/variant

A variant name

Person/name/variant/type

The type of variant of the name

Person/uri

A URI with information about the person.

Person/affiliation

The affiliation of the person within an organization.

Person/affiliation/name

The name of the affiliation of the person.

Person/affiliation/description

A more detailed description of the affiliation of the person.

Person/affiliation/organization

The organization with which the person is affiliated.

Person/identifier/value

An identifier of the person according to an international identifier scheme.

Person/identifier/type

The international identifier scheme for the identifier of the person.

Person/contact

Contact information for the person, including address, phone number, and email.

Organization/name

The name of the organization.

Organization/name/primary

The name of the organization (if it has variants).

Organization/name/variant/type

The type of variant name of the organization.

Organization/name/variant/primary

The variant name of the organization.

Organization/subdivision

The subdivision of the organization directly involved with the production of the bibliographic item.

Organization/subdivision/primary

The name of the subdivision (if it has variants).

Organization/subdivision/variant/type

The type of variant name of the subdivision.

Organization/subdivision/variant/primary

The variant name of the subdivision.

Organization/uri

A URI with information about the organization.

Organization/name/identifier/value

An identifier of the organization according to an international identifier scheme.

Organization/name/identifier/type

The international identifier scheme for the identifier of the organization.

Organization/name/contact

Contact information for the organization, including address, phone number, and email.

Organization/abbreviation

Abbreviation under which the organization is known.

=== Title

==== Preferred form

• Source: 4.2; 7.3

• Serialisation: bibitem/title

The wording and spelling of the title should be given in the form in which it appears in the cited information resource (see ).

A title is serialised as bibitem/title, with its type in title/@type and its language and script in the language attributes (see ).

Additional information such as the format or date may be provided, if necessary.

A work identifier (e.g., ISAN) may be added after the title, preceded by the identifier type. The identifier should be used if the title alone does not uniquely identify the cited resource.

==== Alternative forms

If alternative forms of title are given in the cited information source, the most prominent should be used. If several forms are given equal prominence, the form used should be one of the following, listed in order of preference:

1. a title in the language and script mainly used in the cited information resource;

2. the title given prominence by the order of titles or the layout;

3. a title in the language of the main target audience; and

4. other forms of title.

A resource can have multiple titles, including alternative titles and subtitles. Titles are classified through title@type; the model allows for “alternative”, “subtitle”, “unofficial” (i.e. a title that has become prevalent but has never been the official/intended title of the resource), “original” (including the source language title of a translated resource), and “main” (for the proper title or titles of the resource). If a title type is not given, it is assumed to be a main title.

==== Popular or original title

If the information resource used is widely known by, or was originally issued under, a title different from that of the cited version, the alternative may also be given. The alternative title shall be placed in brackets immediately following the original title.

A popular title is serialised as title@type = "unofficial", and the source language title of a translated resource as title@type = "original".

==== Long title

A title that is inconveniently long may be shortened by the omission of some words, but initial words, other than the definite or indefinite article, should not be omitted, nor should the sense be altered. The omitted words, except for omitted articles, should be replaced by an ellipsis indicated by three dots (for abbreviation of periodical titles, see ).

Even the truncated title should enable unique identification of the cited information resource. Generic titles, even if long, should not be truncated.

It is not possible to specify a generic logic or method for truncation, which diminishes the usefulness of title shortening.

==== Additional title parts

Subtitles shall be included if they are required for the identification of the cited information resource (see ), otherwise they may be omitted. A subtitle is serialised as title@type = "subtitle".

If the cited information resource consists of two or more parts, the full title of the part used shall be included.

==== Ambiguous or incorrect title

Elucidation of a title that is ambiguous or fails to indicate clearly the content of the cited information resource may be supplied in brackets after the title. If a title contains an obscure abbreviation, the full form, if known, may be supplied.

==== No title

If there is no clear title in the cited information resource or in embedded / external metadata, a popular or traditional title may be given, if one exists. Otherwise, a title should be supplied in brackets, giving the following information:

1. the special subject matter; and

2. the nature of the item (e.g., pamphlet, map, plan, drawing, index, engraving, photograph, dataset).

“Untitled” can be treated as a main title.

==== Translated title

A translation of the title (and/or its Romanized form) of the cited information resource in the language of the publication may be supplied immediately after the original title.

A translated title is serialised as an additional bibitem/title with its language and script attributes.

If both translated title and its Romanized form are given, a semicolon may be used as a delimiter between them.

==== Titles of translated works

The original title of a work, or its translation or Romanized form, may be supplied immediately after the title used in the cited information resource.

==== Representation

The title element can have a key role in enabling the users to locate the cited information resource. Therefore, the title may be emphasized.

For translated and Romanized titles placed in brackets, the emphasis should be used in order to differentiate between this title-derived information and other supplementary information also in brackets.

Titles in scripts that are not traditionally emphasized, should not be emphasized.

=== Component parts

==== General

• Source: 7.4

• Serialisation: bibitem/relation[type=includedIn|includes], bibitem/extent, relation/locality

Citable information resources can be component parts of host resources, other citable information resources.

Component parts include articles in periodicals, newspapers and Web sites, individual tracks (and their lyrics) in LPs and CDs, poems and short stories published in books, episodes in TV series, conference presentations published in proceedings, and so on.

A component part is serialised as a relation of the item within its host item (see ); the position of the component part within the host item is given in relation/locality, and its extent in bibitem/extent.

==== Representation

===== General

If the host information resource is not a serial, a connector phrase. “In:” or equivalent in the language of the publication should precede the title of the host item.

Serials include e.g. periodicals, journals, newspapers, annuals etc.

The method of representation is given below, where {hostItemTitle} is the title of the host item.

It is also possible to provide a more complete description of the host resource, where hostResourceReference is the reference to the host item:

Locality information within the host item may be provided, where locality is the location of the component part within the host item:

===== Serials

The use of the connector phrase is not needed for citations which are contributions from journals. It is optional for other types of serials.

===== Emphasis for locatability

Reference should enable the users to locate the cited information resource, no matter how granular it is (for instance, a component part of a component part, such as an image in a journal article).

• If the cited component part can be directly discoverable, emphasis should be placed on its title.

  Belle de Jour [article] [online]. In: Wikipedia. Available from: Wikipedia (distributor), https://en.wikipedia.org/wiki/Belle_de_Jour_%28film%29. [viewed 2018-08-24].

  CLARK, Carol. On the threshold of a brave new world. In: Blueprint of the body [online]. CNN.com, 2000 In-depth specials. Archived copy available from: Internet Archive (distributor), https://web.archive.org/web/20010223233730/http://www.cnn.com/SPECIALS/2000/genome/story/overview/ [viewed 2018-02-14].

  DÜRER, Albrecht. Feldhase [still image] [online]. In: ART IN WORDS. Wien: ARTinWORDS, 2018. Available from: https://artinwords.de/albrecht-duerer-feldhase-1502/. [viewed 2018-07-23].

  AYMARD, M. (ed.), Dutch capitalism and world capitalism. (Studies in Modern Capitalism). New York: Cambridge University Press, 1982, pp. 78–96.

  MADDOX, Brenda. DNA’s double helix: 60 years since life’s deep molecular secrets was discovered. [online]. The Guardian, 2013-02-22. ISSN 1756-3224.

• If the cited component part cannot be retrieved but the host resource can be retrieved, emphasis should be placed on the host item’s title.

  MADDOX, Brenda. DNA’s double helix: 60 years since life’s deep molecular secrets was discovered. [online]. The Guardian, 2013-02-22. ISSN 0261-3077.

  MANN, T. Letter to Hermann Hesse, 1951-10-14. In: WINSTON, R. and C. WINSTON (ed.). The Letters of Thomas Mann, 1889–1955. Vol. II 1942–1955, p. 630–632. London: Secker & Warburg, 1970.

  COHEN, Leonard. Suzanne [lyrics]. In: COHEN, Leonard. Songs of Leonard Cohen. [audio disc]. Columbia records, 1968.

  COHEN, Leonard. Suzanne [lyrics]. [online]. Available from: . [viewed 2019-12-10].

When citing an article in a digital journal where the article can be found directly, the emphasis is placed on the article’s title, since the article is the most granular locatable resource.

When citing a newspaper article from a printed manifestation, the emphasis is placed on the newspaper’s title (the host item) since it facilitates location of the article (the cited resource).

=== Formats and resource type categorization

==== Formats

• Source: 4.4; 7.5

• Serialisation: bibitem/medium

An information resource with identical content can be published in multiple formats, potentially manifesting via multiple types of media.

While some formats are labelled as “editions”, this document uses the term “edition” specifically to indicate content revision.

In specifying format, it is recommended not to specify the reading or interpreting device used. It is generally unnecessary to indicate the actual digital file format.

differentiates between the content type of the item, the media type (which includes file format for electronic media), the carrier type, the size of the item (which can include duration, and electronic file size, as well as physical medium size), and for maps, the scale of the item. These are all modelled as elements of the item medium; the media type or file format is modelled as medium/form.

There is no fixed vocabulary modelled for any of the elements of medium (although suggests closed Resource Description and Access (RDA) vocabularies for medium/content and medium/carrier, and IANA Media Type is best practice for medium/form for electronic resources. As allows, there may be instances where the medium descriptions need to be discursive.

suggests specifying detailed information about the type of electronic information resource (e.g. “statistical dataset”), which cannot be represented in the type of bibliographic item (bibitem/@type), and is likely to be to specific for the RDA triple of content, carrier and form. The medium/genre element is intended to convey this kind of information, and can be used more generally as a well-recognised subclass of bibliographic type or content.

Information about alternate formats should be modelled as related bibliographic items (through the “instance” document relation), each with their own medium information.

The size and scale of the item (medium/size, medium/scale) usually contains numerical information, but is modelled as a string, given the complications of how size and scale can be described, particularly for cartographic material.

==== Resource type categorization

===== General

Content, media and carrier type should be specified if they have a significant impact on the intellectual content, look and feel, or usage of the cited information resource.

The information should be taken from the resource if it is provided. Otherwise, an external reference may be used when necessary for identification purposes, such as:

• IANA Media Types list

  ;

• Resource Description and Access (RDA) content types

  , media types
  , and carrier types
  .

Resource type description does not always provide added value for citations.

===== Content type

Content type should be provided when it is required for the identification of the cited work. For instance, there may be several related interrelated works which share the same title.

===== Carrier type

Digital information resources can be fragile. In order to preserve them, it will be necessary to transfer them to new carriers. For instance, audio tapes last only up to 30 years, but their lifespan can be even shorter if the tapes are not stored correctly. And even if the carrier itself is still fine, finding a player for an outdated carrier type (e.g. a C-cassette deck) may be difficult.

Since carrier type information will become outdated, carrier type should not be specified if the same version of the information resource is known to be available on multiple carriers (e.g. LP, C-cassette, CD) or if the information used is available even on versions published on other carriers.

Carrier type should be provided if the carrier type is relevant for the citation.

==== Digital file formats

The digital file format of the cited online electronic information resource should be specified. This is especially the case when the carrier type is not provided or if the carrier type does not provide sufficient detail about the file format.

Identification of digital file formats may utilize the following information:

• file extension of the information resource;

• ;

• well-known format designations like the Unique Identifier.

Both IANA Media Type and file extension may be insufficient if there are several incompatible versions of the same file format with significantly different characteristics. In such cases, different rendering tools may be required.

If the format of the cited resource is uncommon, sufficient technical metadata or a link to such metadata should be provided. In cases where the cited resource is from an archive with a digital preservation program, this may be covered by the archive.

A file format specification can be accompanied by one or more applications capable of rendering resources in the format, and for provenance purposes the application with which the information resource was created should sometimes be provided. However, including technical metadata in references is not very practical since it changes over time. If file format is specified using the PRONOM identifier, the identifier can be used to find up to date information about the format, including rendering tools, from the PRONOM database.

If the format is widely known and supported (e.g., HTML), the RDA term “online resource” or just “online” may be used instead of the file format designation. Type of the resource may also be specified.

The specific type of electronic information resource may also be specified, if it is not clear from the context. The level of specificity may be chosen, as appropriate.

Information about other file formats in which the cited information resource is available (e.g., a PDF version of an HTML document which was originally used) may be provided in addition to the cited resource. Information about alternative formats should be clearly separated from the information pertaining to the file format used, and it should be preceded by an appropriate phrase, including the file format and the online host if necessary.

If an alternative file format is included in the reference, a check should be made to verify that the alternative resource contains the cited information. Digital surrogates of printed resources can differ significantly from the original documents.

==== Migrated medium

If the information resource used has been migrated and the original resource is no longer accessible, the migrated resource may be used if it contains the cited information. A note about the original version may be added.

=== Edition and version

==== Editions

• Source: 4.5; 7.6

• Serialisation: bibitem/edition, bibitem/version, bibitem/date

Editions are used to differentiate significant content revisions, including draft stages.

Editions are modelled as free text, given the variety of ways they can be represented. For that reason, the full word “edition” needs to appear after any ordinal. The optional number attribute can be used to provide a numeric equivalent.

An edition number shall be provided if the information resource is of a second edition or higher. The edition should be specified in the terms, including symbols, used in the cited information resource. The word for “edition” or equivalent term in the language of the cited information resource may be abbreviated.

If there is a reasonable expectation of forthcoming editions, an edition number should be provided even when the cited information resource is a first edition.

If a program or research dataset cited is updated often, edition (or version) should be included only if necessary, since the information provided becomes outdated very quickly, and locating and/or running the indicated version is likely to be difficult. An edition number is required if other (earlier or later) editions of the program or research dataset are not applicable.

If an edition (or version) number is provided for programs and research datasets, a first edition is not an exception. The first official release of a program may not be the first version that has been available for the public. Edition is also important since hardware and software requirements are dependent on version information.

Some information resources may be appropriately cited prior to formal publication. For instance, it is possible to cite a preprint of a scientific article before it is published, or a standard when it is still under preparation (see ).

Citing an unfinished information resource shall follow the guidelines provided either within the information resource or by the publisher. Draft stage indication (such as “Preprint” or “work in progress” shall be incorporated into the reference according to such guidance.

Reprints should also be specified as such, with the phrase “Reprint” or equivalent after the title.

==== Audience differentiated content

If an information resource is published in different languages in different countries, the country of publication should be given after the title in the language used in the source and separated from the main title with a colon.

If the edition of the cited resource shown to a user depends on his or her geographical location, a note indicating the edition used should be provided. If the content of a resource can be personified with e.g., cookies or some other technique, citation should be based on a non-personified basic version of the page.

==== Version

If the cited information resources has been updated or expanded, it usually has a version statement. The number and other designation for the version should be recorded in the terms given in the cited information resource.

Words such as “version”, “level”, “update”, as well as phrases such as “revised”, “third update” or “training level” are often used to indicate version changes.

The version element can be used to give more detailed information about resource versions, including version numbers (drafts) and version dates (which are differentiated from date, since they are often used as version identifiers).

As described in , dates when a resource was last accessed are modelled as date[@type = "accessed"]; dates when a resource was last updated are modelled as date[@type = "updated"].

If a change in a file format version has an impact on the usage (for instance, different tool is needed to render the resource), the version information should be provided.

Personified Web resources should not be used for referencing, since other users may not be able to verify whether the resource does contain or did contain the cited information.

If a specific version of an information resource is cited, the version should be indicated by giving the name and/or number of the release, update or equivalent, and, if necessary, the year, with (if necessary) month, day and time (see ), when that version was created or updated. It is also possible to specify the date when the information resource was viewed or accessed.

Guidelines for citing versions of Web pages and their component parts are described in detail in .

=== Date

==== General

• Source: 7.7

• Serialisation: bibitem/date, date/@type, date/on|from|to, date/@text

The date of publication, manufacture, distribution, execution, transmission, etc., should be cited if necessary. For physical items such as printed books, providing the year is often sufficient to facilitate locatability, but for intangible works (e.g., Internet documents, serials, research datasets and patents), the explicit date and time of the last modification and/or date and time the document was viewed may be necessary.

If appropriate, date may be expressed as shown in the cited information resource or external metadata describing the resource, e.g., the name of a season or a university term.

The format of the date information should be based on representations provided in the ISO 8601 series if the information should be machine readable and understandable.

The ISO 8601 series specifies representations of date and time information for interchangeable purposes, that are sometimes neither familiar nor ideal for human users. For instance, an open time interval intended for human reading is usually represented using an en-dash (e.g. 1990–) or hyphen (e.g. 1990-), whereas its standard-based representation utilizes a solidus and full stops (e.g. 1990/..). In these cases, the human readable form is preferred.

A date is serialised as date/@type (the lifecycle phase, from the BibliographicDateType vocabulary) with either the point date date/on, or the range date/from and date/to, each in the ISO 8601 format. A date that cannot be expressed as a Gregorian date is serialised as text in date/@text.

==== Model

For a point date, only date/on is populated. For a date range, date/from and date/to are populated. If the range is ongoing, the date/to element is left out.

The date for date/on, date/from and date/to is by default given as a Gregorian date/time, following the conventions of ; in an XML serialisation, they would be expected to follow the XML Schema xs:datetime representation (). However bibliographic dates often contain information that cannot be restricted to an ISO 8601 date (dates in other calendars, date corrections; other subdivisions of the year; approximate date or no date.) For that reason, the model allows dates to be expressed in text instead of as an ISO 8601 date (date/on@text, date/from@text, date/to@text), or as well as an ISO 8601 date if a Gregorian date is still applicable. If the textual value of a date is a date range, and ISO 8601 dates are still to be provided for the start and end of the date range, then the textual value is repeated for date/from@text and date/to@text.

Date corrections may also involve indications of provenance ().

The date of copyright is associated expressly with the copyright holder, in bibitem/copyright/date/from. Unlike other bibliographic dates, the copyright date can only have a year value.

===== Date types

Dates are associated with specific phases of the production of a bibliographic item, and those phases are named through date@type. The default type used in bibliography is the date of publication (date[@type = "published"]), but the “created”, “updated”, “unchanged”, and “copied” dates in particular are often given where they differ from the date of publication. The date of transmission is used instead of the date of publication for a broadcast.

The repertoire of date types is expanded to deal with the particular requirements of standards:

Normal bibliographic dates:

created

Date a resource was created.

published

Date a resource was published.

accessed

Date a resource was last accessed by the bibliographer; routinely used for online publications. (Unlike in , no distinction is made between “viewed” and “accessed” based on whether the resource is human-readable or machine-readable.)

updated

Date a resource was updated and republished. Used for new editions.

unchanged

Date a resource was reprinted, i.e. republished without any changes.

circulated

Date a draft version of a resource was circulated. (Drafts by definition are not published.)

transmitted

Date a resource was broadcast.

copied

Date a resource was physically copied, or recreated without any substantial change in content (allowing for change in medium).

adapted

Date a resource was adapted for a new purpose or audience, with some change in content (includes translation).

Standards bibliographic dates:

implemented

Date a standard takes effect.

obsoleted

Date a standard becomes no longer in effect.

confirmed

Date a standard is renewed to take effect.

issued

Date a standard is authorised for publication.

The date of publication is taken to be the date when the current version of the resource was published, unless a separate “unchanged”, “updated” or “copied” date is given; in that case, the date of publication is assumed to be the date of initial publication.

The date of creation is taken to apply to the work instantiated by the resource, rather than to a specific format or edition. The latter are represented by the date of adaptation and of update, respectively.

==== Representation

===== Year

The year should usually be given as it appears in the cited information resource, however, years written as words in the cited information resource should be given as Arabic numerals in the reference.

If the cited information resource bears a date from a calendar system other than the Common Era (CE), the date should be given together with its equivalent in the CE calendar system supplied as associated information.

===== Season, month, day, time

If a machine readable date is required, it should be provided in a format compliant with the ISO 8601 series, including, if necessary, time. In that case, the time zone may also be provided.

The precise date may be provided with the season or month spelled out, either in full or abbreviated, and with day, month, year and time given in the order and language used in the cited resource (see ). If so, an ISO 8601 equivalent of the date may also be provided as associated information after the date from the cited resource.

====== Uncertain, unknown, approximate or missing date

If the year of publication is not given but can be ascertained, it should be supplied. If the date is uncertain, an approximate date, followed by a question mark should be provided.

The character X may be used in place of one or more rightmost digits to indicate that the value of that digit is unspecified. Approximate date may be specified with tilde (“~”), uncertain and approximate date with percent character (“%”).

If no date can be given, the Latin phrase “s. a.”, English phrase “no date” (or “n. d.”), or an equivalent in the language of the publication may be supplied.

====== Date ranges

Date ranges should be created using the Unicode “hyphen-minus” as the separator character. Open ranges are acceptable.

If the date range has to be machine readable, it may be expressed as specified in ISO 8601. The standard uses solidus (“/”) as the separator character in ranges.

===== Date of publication

The year (and, if necessary, full date or date and time) on which the cited information resource was published should be given.

In a reference to the whole or a limited run of any item issued in parts over a period of years, the years of publication of the first part issued (which is not necessarily numerically the first part) and the part last issued should be given.

If the run is in progress, the year of the first item should be provided, followed by a hyphen-minus. The numeration and pagination (see ), if any, should also be given.

===== Date of citation

The full date (and, if necessary, time) on which the cited information resource was viewed should be given if:

1. The cited information resource is dynamic, and it is necessary to indicate the exact version used with the date.

2. There is no other date information to be found either in the cited information resource or in the related documentation such as external metadata.

The date of citation should be preceded by an appropriate term such as “viewed” and “accessed” (for human and machine-readable information resources, respectively) or equivalent terms in the language of the publication.

The date the cited resource was viewed should follow the link used. Even a single citation may contain several links to different copies or versions of the cited information resource, and it may be important to indicate when the functionality of these links was checked.

===== Surrogate dates

In a reference to a reprint, facsimile, digital surrogate or other copy of the original information resource, the date of publication of that edition or, for a copy, the date of the original, should be given, along with the date of reprinting or copying.

===== Incorrect dates

If the date given on the cited information resource is known to be incorrect, the correct year should also be supplied in brackets after the date in the cited resource.

If the date on the cited information resource is known to be different from the actual publication date, both dates may be given; the actual date in brackets after the date in the cited resource.

=== Production information

• Source: 4.6; 7.8

• Serialisation: bibitem/place, bibitem/contributor[role@type = "publisher"], bibitem/contributor[role@type = "distributor"]

==== Roles

===== General

The organizations or people responsible are likely to be variously described in different kinds of resources. The order of preference should be:

1. publisher or production company;

2. distributor or issuing body or online host;

3. printer or manufacturer;

4. sponsoring body.

The publisher is a subclass of creator (). Of the classes of creator that can be used as the publisher, “publisher or production company” and “printer or manufacturer” have bibitem/contributor/role@type = "publisher"; “distributor or issuing body” and “sponsoring body” have bibitem/contributor/role@type = "distributor". allows that distributors, issuers and sponsors can be named in addition to publishers.

===== Publisher

The organization or person shown most prominently in the cited information resource (see ) as responsible for the production of the resource should be given as publisher.

Publishers’ names may be abbreviated by omitting details not essential for identification.

A public identity identifier (ISNI or other) may be added after the publisher’s name, preceded by the identifier type. The identifier should be used if the name alone does not uniquely identify the publisher or if the publisher has changed name.

If a publisher is unknown, an indication such as “publisher not known”, or an equivalent abbreviation such as the Latin “s.n.” may be provided.

===== Multiple roles

If the names of two or more organizations appear equally prominently in the same role, only one need to be given; the other names may be omitted if they do not provide added value.

===== Distributor, issuing body or online host

The name of the distributor, issuing body, or online host should be given in the citation if appropriate and known.

Any restrictions on access to a resource imposed by an online host (e.g. geographical restrictions) should be indicated in bibitem/note.

A well-known acronym may be used, and a translation of the name may be given.

If the online host is different than the publisher, both should be given.

If two copies of the information resource have been used, both should be cited if it is not known whether the copies are technically identical. If they are different or if they have different access restrictions, both shall be cited.

Online hosts may have technical requirements which have an impact on the usage of information resources on their sites. Site specific hardware and software requirements should not be included in references, since such requirements do change often.

If an online host limits access to information resources permanently on e.g., geographical basis, the access barrier should be indicated.

===== Printer or manufacturer

The name of the printer and manufacturer, if applicable and known, should be provided in the citation.

===== Sponsor

Sponsors may have many roles, including funder, contributor, distributor or online host.

The name and, if appropriate and known, the role of a sponsoring body may be given after the publisher’s information.

If the publisher is also a sponsor, sponsorship information can be shown with the publisher.

A publication that is published on behalf of another entity can be described after the publisher.

==== Place

===== General

The place of production should be given as it appears most prominently in the cited information resource. If the place is known but not mentioned in the cited resource, it may be supplied.

For well-known places, the place name alone may be given. If there is another, less well-known place with the same name or if the place is less known, one or more qualifiers should be added (e.g., county, department, state, province, country) to the place name to avoid confusion and to aid correct localization.

Strictly speaking the place of production is an attribute of the publisher, and it can in fact be modelled as bibitem/contributor[role@type = "publisher"]/organization/address/city. However, it is modelled as a top-level attribute of the bibliographic item to allow the place to be extracted by consumers of the serialisation more readily, to allow multiple locations to be associated with the publisher, and to allow only one of the places of the publisher to be associated with the production of the item.

A URI in a geographical registry may optionally be given as place/@uri. A region within which the place is located may be given as place/@region, for disambiguation purposes; using a ISO 3166 code for a country or subnational region is recommended.

===== Place codes

Names of countries and their subdivisions may be encoded in accordance with and .

===== Place identifiers

An identifier may be added after the place name. The identifier may be used instead of a qualifier if the name alone does not uniquely identify the place.

===== Multiple places

If two or more place names appear equally prominently in association with one publisher, distributor, producer, etc., all names after the first in the cited information resource may be omitted.

===== Unknown place

If a publisher is given and no place is known for them, an indication such as “place not known” or a similar phrase in the language of the publication, or an equivalent abbreviation such as the Latin “s. l.” or the English “n.p.” may be provided.

=== Numeration and pagination

==== General

• Source: 4.7; 7.9

• Serialisation: bibitem/extent

The components of the cited information resource should be referred in the terms used in the resource, their order following the principle that the larger precedes the smaller.

Components of an item are modelled under the bibitem/extent element as one or more “localities”; these consist of a locality type (bibitem/extent@type), a locality value (bibitem/extent/referenceFrom), and optionally a locality range endpoint (bibitem/extent/referenceTo). The extent gives the part of the host item being cited, if this is a contained item (), such as a journal article or book chapter (e.g. “pp. 25-45”).

Some electronic formats such as PDF may retain the pagination of the original (printed) resource, but some other formats may not have the pagination. If so, logical components of the cited information resource should be used.

A hierarchical locality is expressed by repeating the extent element with the different applicable locality types; these should be given in hierarchical order. However localities in an extent may be complimentary instead (e.g. “xiii+179 pp” for prefatory and main pagination; “114pp + 3 plates” for plates outside the main page numeration).

Discontinuous ranges are modelled by repeating the extent element for the same locality type.

Extent includes the volume and issues of a journal, if the item is an article in the journal.

If the item is not a component of another item, then the extent gives the locality value for the entire work (e.g. page count: “[7 pp.]”). This is expressed by giving an empty bibitem/extent/referenceFrom, and populating only bibitem/extent/referenceTo (“[7 pp.]” = pp. 1-7).

If a locality is labelled conventionally with a range, but is in fact a single item (as often occurs with volumes and issues; e.g. issue 18–19 is in fact a single issue), the label should be associated with bibitem/extent/referenceFrom rather than being treated as a true range. However the rendering of the item will typically not be any different.

A part of a contained item may be cited, and this has posed difficulty for some citation schemes (e.g. citing a page within a journal article, which itself covers a page range: “Dedes, D. 1993. Ποίηματα του Μαυλανά Ρουμή [Poems by Mevlana Rumi]. Ta Istorika 10.18-19: 3-22. pp. 7-19”.) However, the model copes with this by separating the extent of a citation from the extent of a bibliograhic item, and the extent of a contained item from the extent of a host item. In this instance, “pp. 7-19” would be treated normally as part of the citation, not part of the bibliographic item.

==== Part cited

The reference should identify the part of the information resource that is cited. For a component part, the number and designation of the smallest separately identified component should be given.

If the entire information resource is being cited, its extent may optionally be stated in brackets. The extent should be given, if the cited resource is large.

==== Omission of terms

Previous editions of this document allowed omission of the term “volume” and terms for smaller component parts of serial publications. Typographic means were then used to distinguish the volume number and the part number, with the volume number in bold type and the part number, if required, in parentheses.

In this edition, such practice is no longer allowed. Font styles should not be used to replace or indicate information, because individuals with limited sight or reading disabilities may not be able to distinguish between bold and regular font. Also, generally, a proprietary software, such as Microsoft Word is needed to view bold or italics. In preparation for machine-readable citations, it is important for citation data to be provided in a format that is easily recognizable by a machine, such as a plain text file.

=== Series title and number

==== General

• Source: 4.8; 5.2; 7.10

• Serialisation: bibitem/series

Articles have their journal title modelled as the series title of the article item. However, the journal volume and issue should be modelled as part of the article item extent , rather than as part of the series.

Standards body identifiers for standards are modelled as series numbers.

The collection or fonds of archival material is modelled as a series.

==== Qualifiers for titles

If the title of a serial issued by a society or similar organization is insufficient to identify the cited information resource and the key title is not known, the organization’s name should be added as specified in .

Organization and Place of Publication can be added to series titles for disambiguation. Organization in this instance is only modelled as a strings, rather than as a detailed object (cf. ). The formattedTitle element is used to enforce the formatting of the series title and its qualifiers.

The title may also by qualified by the place of publication or ISSN in order to support the identification of the serial.

==== Key title and the abbreviated key title

Serials with the same title should be distinguished from each other with the key title. These titles are available from the ISSN Portal

Abbreviated key title is a specific title established by the ISSN network by replacing each significant word of a key title with the corresponding abbreviation from the List of Title Word Abbreviations

The abbreviated key title should be used when it appears in the cited information resource (see ). The ISSN should be included with the abbreviated title (see ).

If there is any risk of ambiguity, the title shall be given in full.

==== Earlier titles

If the title of a cited serial has changed, earlier title of the serial should be given with the years when the title was in use.

The current title of a series is given with series@type as “main” (which is the default); alternate titles, including historical titles, are given with series@type as “alt”. Different time spans for historical titles are given with series/dateFrom and series/dateTo.

==== Regular conferences

If regular conferences are held with the same title, the date and place of the conference should be added, in the format used in the cited information resource.

==== Carried title

If the cited information resource also carries the title of a series of which it forms a numbered or unnumbered component, the series title and the numeration within that series may be included in the reference. The series title and number should be recorded as they appear on the cited resource.

The ISSN of the series may be a part of the citation.

A report number should be included.

==== New series

Journals may re-start their series number without changing the journal title. Then it is necessary to differentiate the “old series” and the “new series”, using “new series” or “n.s.” as appropriate.

If a series restarts numbering its volumes, that is typically indicated in citations by giving an ordinal number for the series run, given after the series name: 1st series, 2nd series, 3rd series. More commonly, there are only two series runs for a series, and the second run is indicated as “new series”, or the abbreviation “n.s.”.

• Vizantiiskii Vremmenik (St Petersburg/Petrograd/Leningrad, 1894-1928) and Vizantiiskii Vremmenik (Moscow, 1947-) have the same name, and volume numbering for the Moscow journal restarted at 1; the first two volumes of the new series run were numbered 1/26 and 2/27, to continue the first series. While citations could differentiate the two by city, usual practice is to insert “n.s.” after the journal title for the second series.

• The Journal of the Royal Anthropological Institute was originally published as Man, from 1901-1965. The journal restarted numbering at 1 from 1966-1996 without changing name, and explicitly designating itself as a “new series”. Unlike Vizantiiskii Vremmenik, both series were published in London, so place of publication cannot differentiate between them. While the organization did change name between the two runs (from “Anthropological Institute of Great Britain and Ireland” to “Royal Anthropological Institute of Great Britain and Ireland”), the change is slight; and the series run is preferred to differentiate the two runs.

Subsequent re-starts followed by “new series” would be indicated by the ordinal number of the series preceded by the letters “ser.”, starting from “3rd ser.”

The series number should be recorded as it appears on the cited resource.

=== Identifiers

==== International standard identifiers

• Source: 4.9; 7.11

• Serialisation: bibitem/docidentifier

states that unique international standard identifiers must be included in the bibliography. In practice, identifiers are included at the discretion of the bibliographer: they are routinely included in bibliographic databases, but seldom in references within documents.

If the cited information resource has an international standard identifier (e.g., ISSN, ISBN, ISMN, ISRC, etc.) that uniquely identifies the resource, it may be included in the reference (see ) in a human readable form.

The namespace of an identifier is indicated through its type attribute. The scope of an identifier is introduced to disambiguate the identifier referent, when it is not obvious. So it is introduced where an identifier identifies only a component of the item (e.g. a document included in the item), or a superset of the item (e.g. a series). It can also be used if the identifier identifies a specific instantiation of the document as an entity; in terms of the Functional Requirements for Bibliographic Records , for example, an identifier may identify a work, an expression of the work (e.g. an edition), a manifestation of the work (e.g. a format), or an item of the manifestation (e.g. a single physical copy).

Neither the type nor the scope of the identifier have an associated controlled vocabulary, and one is unlikely to be derived. Identifier type should follow established practice in the domain; most well-known identifiers have a widely used short name or abbreviation.

Cited information resources can have both manifestation and work identifiers. These identifiers have different roles in information retrieval. For instance, a track published on a CD can have ISWC (International Standard Musical Work Code) assigned to the composition, and ISRC (International Standard Recording Code) assigned to a specific recording of the composition. With the latter it is possible to find different copies of the same recording in the Internet or on other carriers; with the former a user can find other recordings or interpretations of the same work, if no manifestations of the cited recording can be located.

Depending on how the information resource is cited, either the work identifier or the manifestation identifier may be the best choice. The former is a good choice if any manifestation of the work will do. Both identifiers may be included in the reference, if appropriate.

ISSN can have multiple roles. There is an agreement to consider it as an identifier for both works and manifestations, because a serial work is also a plan for how a serial will be expressed and manifested. Therefore a serial may only have one expression and one manifestation.

If other metadata elements in the reference do not guarantee unique identification of the cited information resource (for instance, if there are several periodicals sharing the same title) the identifier of the cited information resource should be included.

If the cited information resource has more than one identifier (for instance, a book published in a series may have both ISSN and ISBN), only the one that most precisely identifies the resource used should be included in the reference.

If the cited information resource has both ISBN-10 and ISBN-13, only ISBN-13 should be included in the reference.

If the identifier in the cited information resource is cancelled or invalid, it may be included, preceded by term “Invalid” or “Cancelled” or an appropriate phrase in the language of the publication.

Invalid or cancelled identifiers should be so identified through the identifier scope: the cancellation of the identifier does not impact its namespace, but its connnection to a referent.

If the correct identifier is known, it should be used in the citation. The invalid or cancelled identifier may also be included in brackets, preceded by term “Invalid” or “Cancelled” or an appropriate phrase in the language of the publication.

Standards body identifiers for standards may be modelled as series numbers or as identifiers (). We recommend that the canonical identifier issued for a standard by a standards body be encoded as an identifier; other identifiers assigned to the same document may be better identified with a series title, as a named and numbered grouping of documents.

Identifiers can be expressed as HTTP URIs may be so expressed, without changing the identifier type. A DOI for example may be expressed as either the DOI proper, or as a doi.org URI.

==== Persistent identifiers

If the cited information resource has a persistent identifier (PID) such as ARK, DOI name, Handle or URN, it shall be included in the reference.

Actionable PIDs should be expressed as hyperlinks (HTTP URIs). Un-actionable versions of persistent identifiers such as doi:10.3886/ICPSR31521.v1 or URN:NBN:fi:au:cn:39756A should not be used.

If the PID is present in the reference as a hyperlink, the network location (URL) of the cited information resource shall not be used.

If the cited Web resource has both a traditional identifier and a persistent identifier based on it (for instance ISBN and URN:ISBN or ISBN-A), both should be included in the reference.

If the cited resource is a component part, the traditional identifier of the host information resource should be included in addition to the persistent identifiers of the cited information resource. For instance, a journal article may have a DOI name, and the journal itself an ISSN.

If the cited information resource has more than one persistent identifier, they may all be included in the reference. Each PID facilitates access to a different copy of the cited resource, and these copies may have different access rights (for instance, one provides a link to a copy in a system maintained by the publisher, another to an open archive hosted by a university). If only one PID is included, it should provide a link to a copy that does not have access limitations.

Persistent identifiers often resolve to URLs of landing pages, which contain links to the actual resources. In such cases, it is sometimes not obvious if the PID identifies the cited resource or its landing page (or both), but such PIDs should nevertheless always be included in references, since they facilitate access to the cited resources.

=== Availability and location

==== Model

• Source: 4.10; 5.10; 7.12

• Serialisation: bibitem/accessLocation, bibitem/uri

Unpublished items with a limited number of copies that are held at a repository should be referenced through bibitem/accessLocation, which gives the repository that the cited copy is held in, and the locator used by the repository. The repository and locator are not differentiated, and are treated as a single string. The bibitem/accessLocation element is also used for online locations of digital resources, where a single URL (bibitem/uri) is not sufficient, and either a navigation path, or an in-document object identifier is used.

An archival URI can be differentiated from an original URI for a resource, through using “archival” as the uri/@type. Alternatively, the archival copy and the original copy may be treated as related items (through an “instance” relation), although that is typically not necessary.

Persistent identifiers and other identifiers expressed as actionable URIs should still be represented as bibitem/docidentifier rather than bibitem/uri, so long as they are managed in a distinct namespace. That applies to such identifiers as DOI, ARK, Handle, ISSN, etc.

==== Physical location

===== General

If only a single or a limited number of copies of an information resource such as artwork or archival material are known to exist, or if a specific item is referenced, the place name shall be provided in a citation.

===== Place and repository

“At:” or an equivalent phrase in the language of the publication should be given before the place name.

The method of representation is given below, where {placeName} is the name of the place.

The language of the information resource shall be used in representing the place name.

The place name should be qualified if necessary (see ).

If the cited resource is held within a repository (within for instance, an item or a collection), the name of the repository that it belongs to and its locator (such as the call number, locator or identifier) in the form used by the repository.

It is possible to cite surrogates of information resources that no longer exist, such as temporary or destroyed work.

Examples of full references are provided below.

==== Network location and access

===== General

If the cited information resource is available in the Internet, a hyperlink such as HTTP URI, shall be provided to allow retrieval of the resource.

Internet addresses are often unreliable and therefore, URLs should be used with caution. Recommended practice is to use URIs based on persistent identifiers (PIDs), permalinks or a URI of an archived copy of the resource. If these approaches are not possible or practical, URL may be used, but with caution (see below).

The elements of URI as specified in , URI Generic syntax (e.g., scheme, authority, path, query and fragment) should be transcribed with the same punctuation, upper-case and lower-case letters as given in the cited information resource.

In URIs, the scheme and host are case-insensitive and normally provided in lower case; all other URI components may be case-sensitive and should be treated as such. The name of the service may be added before the URI. Some services (for instance, Google Maps) apply internal URI syntaxes which differ from the URI syntax defined in . In such case, the syntax of the service may be used.

===== Network access

To indicate network access location of the cited information resource, a phrase equivalent to “Available from:”, in the language of the publication, should be given before the URI.

The following indicates the method of representation, where {URI} indicates expression of the URI:

Name of the service (for instance, an institutional repository) holding the cited resource may be included. Below {URI} indicates expression of the URI and {service-name} the name of the service:

Specific features of the cited information resource should be provided, if they have an impact on its use:

===== Related locations

If the cited information resource is for instance a program, it may be useful to include in the citation a link to additional information about the resource. A phrase “See also:” or equivalent in the language of the publication should be prefixed to a URI.

===== Archival access

Web archives make it possible to cite Web sites or individual pages that have been lost, and past versions of resources which are still available but in an edited form.

A link to an archived copy of the cited information resource may be added to the citation even if the cited information resource is static and there is a reason to believe that it will be available on the Web for a long term.

===== Specific versions

If the cited Web page is dynamic and a specific version is used, an archived version of the resource shall be cited. If an accurate archived copy of the page does not exist, one shall be made using an archiving facility if possible, e.g. The Internet Archive.

The citation shall contain the name of the cited resource and the dates it has been used and archived.

===== Nonspecific versions

A link to an archived copy is not required if the cited information resource is dynamic, but no specific version of the resource is cited.

===== Representing lengthy locations

If the location information needs to be broken at the end of a line, the break should be made after a single or double solidus or a hyphen, but before other punctuation marks.

===== Providing navigation path instead of URI

If the cited information resource has long and/or complex location information, the navigation path to the resource may be provided instead of the full URI.

Navigation path may have 1-n levels, indicated in an appropriate manner.

===== Providing PID

HTTP URI based on the actionable persistent identifier (PID) assigned to the cited information resource shall be provided instead of the actual location (URL) of the resource. If the location changes, it may not be redirected to the new URL, whereas the PID will most likely be resolved to the new location or locations.

===== Outdated location information

Due to HTTP redirection, outdated URLs may still provide access to the cited information resource. However, all URLs should be up to date by the time citations are made, since URLs that require redirection are likely to be more short-lived than the current ones.

If the cited information resource has already disappeared from the Web or if the current version has already changed by the time the citation is written, the original URL and the retrieval date of the cited resource should be included, alongside the URI of the Web archive copy of the resource.

Non-operational URLs shall be preceded by a phrase “Originally available from:”, “Previously available from:” or equivalent phrases in the language of the publication.

If there is no Web archive copy of the cited Web resource, or if the existing copy or copies are not identical to the cited information resource, the Web resource shall not be used for reference, even if the author has a local copy of the resource. Proving the authenticity of such a copy may prove difficult.

==== Permanent links

Permanent link systems support issuing of URIs “intended to remain unchanged for many years into the future” (). They work by issue of persistent URL or pattern for persistent URLs, which redirects to the URL of the requested web resource. Widely used permanent link systems include PURL (Persistent Uniform Resource Locator, see ) and .

URL shorteners () (providing URL shortening and redirection services), which may be seen as permanent link systems, should not be used for referencing, because:

• it is not possible to know how long a specific short URL will be supported.

• short URLs do not have a persistence related component unlike for example PURLs. The only role short URLs do have is to make URLs shorter in order to simplify their usage.

• as a rule, short URLs do not function in Web archives, which makes it almost impossible to find the cited content after the short URL has stopped working.

Functionality supported by these systems is simpler than in PID systems and unlike PIDs, permanent links are usually technology (HTTP) dependent.

PIDs are expressed as HTTP URIs with resolver URL, but at least URNs could be actionable as such.

Permanent links may be used if:

• The resource has a registered permanent link (and it is known).

• There is a reason to believe that the assigned permanent link will be managed. However, it may be difficult to find out the referent’s commitment to administer the permanent link in the long term.

• The permanent link system itself is well managed and sustainable.

If the permanent link system is discontinued due to, e.g., technical or financial issues, all URIs based on the system may fail at once. Therefore, references should contain also a link to an archived copy of the page if one exists, and the date/time the cited resource has been viewed.

It can be difficult to find out if the cited information resource does have a permanent link. It is usually not embedded in the resource itself, and the permalink is not be part of the URL to which the permalink is redirected.

==== Access restrictions

The cited information resource may be located behind a paywall or there may be legal restrictions that significantly restrict access the use of the resource. Such restrictions should be specified after the URI, with the date the restriction will end (if applicable). The name of the commercial access provider may be given before the URI.

A note on the access limitation may be given after the URI.

If the cited information resource is no longer available online and there is no appropriate Web archive URL to refer to (e.g. due to changing archive interface tools), the reference may be made to the archival metadata including the name of the archive, archiving date and the archive URL.

A full reference example provided below.

=== Surrogate

• Source: 7.14.8

• Serialisation: bibitem/relation[type=manifestationOf|reproductionOf]

A digital surrogate of a work is serialised as a relation from the surrogate to the resource it embodies or reproduces (see ).

If the cited information resource is a work of art, there may be many digital surrogates of it available in the Internet. They may differ significantly from one another, with respect to reproduction fidelity and quality.

If the cited information resource is a surrogate in Internet, HTTP URIs of the surrogates shall be included. If several surrogates of the same work have been used, they should all be included.

The name or nom de plume of the person or organization which uploaded the resource, date and the distributor should be included.

For clarity, the phrase “digital surrogate” in the language of citation may be used to indicate the nature of the copy.

All surrogates used should be included in the reference.

Names of the online hosts and URIs of the surrogates should be provided. Specific features of surrogates such as size may be included when deemed relevant.

=== Relationship

==== General

• Source: 7.14

• Serialisation: bibitem/relation/@type, bibitem/relation/bibitem, relation/locality

A cited information resource is often related with other information resources.

For derivative works, the original information resource should be referenced within the reference to the derivative work.

The original work, if available, should be provided as an embedded reference before the date of citation in the following form, where relationLabel is the relation label of the relationship, and originalWorkReference is the reference to the original work.

In order to minimize confusion, relations are considered the second type of data elements in , so they can be placed accordingly.

An embedded reference to the original work is serialised as relation with the target item in relation/bibitem, the kind of relationship in relation/@type (the vocabulary of ), and, for a part-whole relationship, the extent in relation/locality.

==== Model

DocumentRelation and related models

DocumentRelation/type

The type of document relation described. Different resource types have differnet relations between documents; but inclusion is applicable to most resource types (though there are constraints in which resource types can include which.)

DocumentRelation/bibItem

The bibliographic item to which this bibliographic item is described as related to.

DocumentRelation/bibLocality

The extent of the this bibliographic item which is related to the other bibliographic item as expressed in this instance, provided that it is not the entire bibliographic item that is so related.

The vocabulary of relation types is given in .

==== Translation

Translations are expressions of the original (translated) work, and are usually indexed under the original author’s name. The translator should be considered a subsidiary creator with the role of “translator” or its abbreviation (see ).

If the translated information resource is an exact counterpart of the original information resource, the source and destination languages shall be included if the original and translated title do not clearly indicate them.

The phrase “Translated from …​ to …​” or its equivalent in the language of the publication should be shown as an attribute of the cited information resource (see ).

If the cited information resource has been translated to the language of the publication, the destination language may be omitted.

Both the source and destination language may be omitted if the titles indicate them.

A reference to the original work should be included in the citation reference as a data element if the translation has been made via an intermediate language. The phrase “Translation of:” or equivalent in the language of the publication should be used as the relation label to the translated expression. The phrase “Original work:” should be used to indicate the source.

Information about the original work or the manifestation used as the source may be included even if there is no intermediate language.

If it is uncertain on what original work, expression or manifestation the translation is based on, the source shall not be mentioned in the reference.

A translation is serialised as relation/@type = "translatedFrom" with the original work as the target bibitem, and its inverse as hasTranslation.

==== Commentary

Commentaries are considered expressions instead of new work. They should be indexed under the original author’s name.

A reference to the original work, if available, should be inserted to the citation reference (see ) with the relation label “Commentary on:” or equivalent.

==== Annotations

Annotations are considered expressions instead of new works. They are usually marketed under the original author’s name.

A reference to the original work, if available, should be inserted to the citation reference (see ) with the relation label “Annotation on:” or equivalent.

==== Arranger

Arrangements are considered expression instead of new work, similar to translation. They are usually marketed under the original author’s name.

The reference should indicate the arranger and purpose of arrangement.

The arranger should be considered a subsidiary creator with the role of “arranger” or its abbreviation (see ).

The phrase “Arranged for {purpose}” (where “purpose” is the purpose of arrangement) or its equivalent should be shown as an attribute of the cited information resource (see ).

==== Abridgement

Abridgements are treated as expressions and not original works. They are usually marketed under the original author’s name.

The person who made the abridgement should be considered a subsidiary creator with the role of “adapter” or its abbreviation (see ).

If a reference to the original work is available, it should be inserted in the citation reference as a data element, with the original reference considered as a relation with the current (see ).

The phrase “Abridgement of:” or equivalent should be used as the relation label.

Otherwise, the label “abridgement” should be provided immediately after the title to indicate it is an abridgement of the original.

Translations may be abridged as well. If such translation is cited, the reference should specify differences using the terminology of the cited information resource, translated to the language of the publication.

==== Reviews and critiques

Reviews are new works, indexed under the reviewer’s name.

The phrase “Review of:” or its equivalent may be used to reference the work reviewed.

The reviewer is considered the most prominent creator for a review. The original reference should also be given if available as a relation (see ).

The phrase “Review of:” or equivalent should be used as the relation label.

In some cases, where the title already provides a reference to the original work, the embedded reference to the original work may be omitted.

=== Relation types

bibitem/relation/@type gives the base class of the relation between two bibliographic items.

The base class of relation types is taken from , and represents generic bibliographical requirements, particularly inclusion in a host item (: includedIn, includes) and translation (: translatedFrom, hasTranslation). The class has been expanded substantially to deal with the requirements of standards documents.

The complete vocabulary, with definitions, is maintained in the relaton-models repository as the DocumentRelationType enumeration; it is reproduced here in the model order.

includes

document A includes document B, through a well-defined relation outlined in , and with distinct metadata

includedIn

document A is included in document B, through a well-defined relation outlined in , and with distinct metadata

partOf

document A is included in document B, with no distinct metadata

merges

document A results from a merger of earlier documents B1, B2, B3…​

mergedInto

The main document is one of the documents merged into the later target document.

splits

document A is split off from earlier document B

splitInto

The main document was split into several later target documents.

instanceOf

document A has as an instance document B (e.g. an edition, a format, a single copy)

hasInstance

The main document is a more specific reference to the bibliographic item, and the target document is a more generic reference.

exemplarOf

The main document is a single physical instance of the bibliographic item, which is more generically referred to in the target document. This corresponds to Item in the Functional Requirements for Bibliographic Records (FRBR) model. Inverse of hasExemplar relation.

hasExemplar

The main document is a more generic reference to the bibliographic item, of which the target document is a single physical instance. The main document may be a Work, Expression or Manifestation under the Functional Requirements for Bibliographic Records (FRBR) model.

manifestationOf

The main document is an embodiment in a particular medium of the content of a bibliographic item. which is more generically referred to in the target document. This corresponds to Manifestation in the Functional Requirements for Bibliographic Records (FRBR) model. Inverse of hasManifestation relation. Includes the more specific relations reproductionOf and reprintOf.

hasManifestation

The main document is a more generic reference to the bibliographic item, of which the target document is an embodiment in a particular medium. The main document may be a Work or Expression under the Functional Requirements for Bibliographic Records (FRBR) model.

reproductionOf

The main document presents the same content as the target document, and the main document has been created to reproduce the target document faithfully. The main and target document are not necessarily in the same medium. Inverse of hasReproduction relation. Includes the more specific relation reprintOf.

hasReproduction

The main document presents the same content as the target document, and the target document has been created to reproduce the main document faithfully. The main and target document are not necessarily in the same medium.

reprintOf

The main document presents the same content as the target document, and the main document has been created to reproduce the target document faithfully. The main and target document are both print publications. Inverse of hasReprint relation.

hasReprint

The main document presents the same content as the target document, and the target document has been created to reproduce the main document faithfully. The main and target document are both print publications.

expressionOf

The main document is a particular realisation of the intellectual or artistic content of the bibliographic item, which is more generically referred to in the target document. This corresponds to Expression in the Functional Requirements for Bibliographic Records (FRBR) model. Inverse of hasExpression relation. Includes the more precise translatedFrom, arrangementOf, abridgementOf, and annotationOf.

hasExpression

The main document is a more generic reference to the bibliographic item, whose intellectual or artistic content the target document realises. The main document may be a Work under the Functional Requirements for Bibliographic Records (FRBR) model.

translatedFrom

document A is a translation of document B

hasTranslation

document A has a translation as document B

arrangementOf

The main document has the same intellectual or artistic content as the target document, and the main document has been created to realise that content through different resources than the target document. Is typically understood to involve the realisation of a musical work with different instruments or voices than the original (arrangement). This includes adding parts or an accompaniment. Inverse of hasArrangement relation.

hasArrangement

The main document has the same intellectual or artistic content as the target document, and the target document has been created to realise that content through different resources than the main document. Is typically understood to involve the realisation of a musical work with different instruments than the original.

abridgementOf

The main document presents a subset of the intellectual or artistic content of the target document, but is still intended as a complete work, presenting a shortened version of the main document expression. Inverse of hasAbridgement relation.

hasAbridgement

The main document has a target document which presents a subset of its intellectual or artistic content, but which is still intended as a complete work, presenting a shortened version of the main document expression.

annotationOf

The main document incorporates part or all of the target document, and enhances it with explanatory commentary. Inverse of hasAnnotation relation.

hasAnnotation

The main document has a target document which incorporates part or all of it, and which enhances it with explanatory commentary ot present in the source expression.

draftOf

document A is a draft of document B

hasDraft

document A has as a draft document B

hasPredecessorDraft

The main document is a generic reference to a work (whether published or pre-published), and the target document is a specific pre-publication version preparatory to the main document.

predecessorDraftOf

The main document is a specific pre-publication version of the work represented by the target document (whether the target document published or pre-published), and preparatory to it. Inverse of hasPredecessorDraft relation.

hasSuccessorDraft

The main document is a generic reference to a work (whether published or pre-published), and the target document is a specific pre-publication version revising the main document.

successorDraftOf

The main document is a specific pre-publication version of the work represented by the target document (whether the target document published or pre-published), and revising it. Inverse of hasSuccessorDraft relation.

editionOf

The main document is a specific published version of the work represented by the target document. Inverse of hasEdition relation.

hasEdition

The main document is a generic reference to a work, and the target document is a specific published version of the work.

updates

document A updates document B

updatedBy

document A is updated by document B

derivedFrom

document A is derived from document B

derives

The main document is the original work from which the target document is derived.

describes

document A describes document B

describedBy

document A is described by document B

catalogues

The main document includes a bibliographic record of the target document. Inverse of cataloguedBy relation.

cataloguedBy

The main document has its bibliographic record included in the target document.

hasSuccessor

The main document has ceased fulfilling some function, and the target document has assumed that function in its stead. Typically applies when the main document is a periodical publication which has ceased publication, and the target document is a new periodical publication, designated as the continuation of the main document. Inverse of successorOf relation.

successorOf

The main document has assumed the function of the target document, which has ceased fulfilling that function. Typically applies when the target document is a periodical publication which has ceased publication, and the main document is a new periodical publication, designated as the continuation of the target document.

adaptedFrom

The main document has its intellectual or artistic content derived from the target document, and has modified it to match the requirements of a different medium or audience. In music, an adaptation changes the musical content, whereas an arrangement changes the instrumenal and vocal resources of the work. Inverse of hasAdaptation relation.

hasAdaptation

The main document has been modified in the target document, to present its content matching the requirements of a different medium or audience.

adoptedFrom

document A is adopted by a standards organisation from document B by a different organisation

adoptedAs

The main document has a counterpart in the target document, which derives it content from the main document, and has been adopted in response to the main document. Typically it is an international standard which has had national standard bodies formulate a local counterpart.

reviewOf

The main document is an evaluation of the target document. Inverse of hasReview.

hasReview

The main document is evaluated in the target document.

commentaryOf

The main document provides explanatory commentary on the target document. Unlike annotationOf, the main document does not incorporate the complete text of the target document, and so cannot be read independently of the original. Inverse of hasCommentary.

hasCommentary

The main document is given explanatory commentary in the target document. Unlike hasAnnotation, the target document does not incorporate the complete text of the main document, and so cannot be read independently of the original.

identical

document A is adopted from document B, without any change

nonequivalent

document A is adopted from document B, and has been altered textually significantly

equivalent

document A is adopted from document B, without any significant textual change

related

document A is related to document B in an otherwise unspecified fashion

hasComplement

The main document has additional or contextual information provided in the the target document, to help understand the main document. The target document is complementary to the main document.

complementOf

The main document is complementary or supplementary to the target document, and provides additional or contextual information to help understand the target document. Inverse of hasComplement relation.

obsoletes

document A supersedes document B, and renders it obsolete

obsoletedBy

document A is superseded by document B, and is rendered obsolete

cites

The main document references the target document. Inverse of isCitedIn relation.

isCitedIn

The main document is referenced in the target document.

=== Other elements

==== General

Additional categories of information that may be included are given below. When such information is included in a reference, it should comply with the guidelines provided.

==== Classification

• Source: 4.11

• Serialisation: bibitem/classification

Classification codes may be included, if they have been recorded in the cited information resource. The classification scheme should be indicated if a code is provided.

The classification scheme is given in classification/@type.

==== Keywords

• Source: N/A

• Serialisation: bibitem/keyword

Keywords associated with a bibliographic item may be given as plain strings, or as terms from controlled vocabularies (with the vocabulary identifier and term code). Taxonomic names may also be given as a taxon hierarchy.

==== Size

The size of a printed document may be included in a reference. It may, if appropriate, be given as:

1. the height only, in millimetres,

   210 mm.

2. the height followed by the width, or,

   210 mm × 150 mm.

3. if appropriate, the A-size designation (see ) followed, when necessary, by “landscape”.

   A5 landscape.

If the printed document cited is exceptionally large, the size should be specified.

==== File size

The file size of a cited digital information resource may be given after the URI of the resource. If the file is exceptionally large, the size should be specified.

==== Price

• Source: 4.11

• Serialisation: bibitem/price, bibitem/note

Information about price may be given, as quoted by the publisher.

Price is marked up as a distinct element, which contains the currency code, represented in ISO 4217. The price should be treated as nominal, rather than capturing all possible pricings at a given time.

Availability information for a bibliographic item is treated as a note. This also applies to the availability of an online resource, and any access limitations, but not to licensing, which is encoded separately.

Price and format () availability may be given together.

==== Registered trademark

A registered trademark may be denoted by the symbol “®” next to the trademarked phrase(s) or symbol(s).

==== Rights metadata

• Source: 4.11

• Serialisation: bibitem/license, bibitem/copyright.

Copyright status of the information resource may be provided, if known.

Where available, URIs should be used to identify licenses unambiguously. If a URI is not available, a commonly accepted abbreviation or short name of the license should be used.

Copyright is privileged with a separate element, incorporating copyright date and owner.

If it is certain that no exclusive intellectual property rights apply to the cited information resource, e.g., it is in the public domain or is not subject to copyright, this information may be included in the reference. The phrase “Public domain” or its equivalent in the language of the publication should be used.

This allows the reader to know that the cited information resource may be used freely.

Materials may enter the public domain in one country sooner than they may in another. A work marked to be in the public domain in a jurisdiction may still be under copyright protection in another jurisdiction. Therefore, the jurisdiction where the cited information resource is in the public domain should be indicated whenever necessary.

Some created works are automatically placed in the public domain since they are not copyrightable. The jurisdiction where the cited information resource where this condition applies should be indicated.

If intellectual property rights apply, the copyright date may be cited, preceded by the symbol “©” or “cop”. Alternatively, the English word “copyright” or its equivalent in the language of the citation may be used. If the copyright date is different from that of the publication date, both may be given.

If intellectual property rights apply but there is a well-known license which allows the usage, the acronym of the license should be provided with the version number (if any). Language of the license may be included as the code. The acronym should be preceded by the English word “license” or its equivalent in the language of the citation.

Example references are provided below.

==== Provenance and authenticity

• Source: 4.11

• Serialisation: bibitem/note

The scope of provenance information varies depending on the information resource described. Traditionally, it has been limited to a record of ownership of a work of art, but provenance information may also describe the historical context of the data and its origins, including the workflows and applications used to generate the research data.

If the cited information resource is a Web document, the citation may contain a checksum with which a user can make sure that the document she has retrieved is the one cited.

If the cited resource consists of multiple files, checksums should be calculated for all of them.

Notes on bibliographical provenance, and checksums for digital resources, are both given as bibliographic notes.

Checksum algorithm should precede the checksum value. A standardized algorithm such as SHA3-256 () should be used.

There are free Web services for calculation of checksums.

An example reference is provided below.

===== Provenance of reference information

Information in a reference typically comes from the publisher of a resource, and is included in the resource itself, as described in 4.1.2. 4.1.2 permits information to be supplied by other parties, and differentiated by including it in brackets. This can include both missing information (e.g. “[Untitled]”), and corrections (e.g. “1959 [i.e. 1995]”).

The party supplying emendations to reference information is identified in 4.1.2 as the citer of the reference. However, the citer should be differentiated from the bibliographer, in the case of bibliographies as consumable resources (e.g. as bibliographic databases):

• The publisher provides the primary source information about the resource, typically in the resource itself, but also as a feed of information to a bibliographer.

• The bibliographer compiles a bibliography as a list of references, and uses their judgement to supply information and to impose uniform representation and rendering of the list.

• The citer uses a bibliography within a resource, to identify cited resources, and may further adjust the bibliography they provide within the resource (e.g. for disambiguation, or to disaggregate references included in other references.

The bibliographer and the publisher in the contemporary bibliography universe are more active parties than in the traditional model, which relies exclusively on citer discretion. Bibliographers and publishers are invested in asserting that they are sources of truth about bibliographic resources; and the bibliographies covered by any model of electronic exchange include references provided by bibliographers and publishers, in bibliographic databases—as well as the references consumed and adjusted by citers from bibliographic databases.

So the original reference as obtained from a bibliographic database may need to be differentiated from any adjustments made to the reference by a citer; and the citer adjustments in some contexts may need to be overridden.

If the serialisation of bibliographic items is to reflect such adjustments to content, it will need to indicate the provenance of bibliographic information, and allow multiple versions of the same information, with differing provenance. At the simplest level there will simply be two or more different versions of the reference serialisation, some representing the publisher or bibliographer’s source of truth, and some representing the citer’s adjustments (in brackets). Adjustments made by the bibliographer would also be represented in brackets; this is long-standing routine practice in library catalogues.

A more granular approach would be to permit multiple instances of any element in the reference model, differentiated by provenance: this could be done with a child element (e.g. an XML attribute), or with a nested container. If no provenance is indicated, the information is assumed to be unchanged from the source of truth. A renderer can then choose which alternative to render depending on local requirements:

==== System requirements

In order to facilitate access to electronic resources, references may specify the requirements for accessing and rendering the cited information resources. In practice, providing this information is complicated for some information resource types. For instance, research data sets often require specific software tools and documentation of the content. In such cases, system requirements should be included in references since otherwise re-use of these resources may be difficult or impossible.

Content providers may impose different site-specific technical requirements to access the same information resource. For instance, a user may be required to have a Flash player and support JavaScript in one site, whereas another site may require HTML5 to access the same content. These provider specific system requirements should not be included in references, since the lifetime of such information is often short.

System requirements for programs are application specific and may be provided in a note. The note may include:

• the type of computer or computers on which the program was designed to be used;

• the name of the operating system or systems and its version on which the program will run;

• the middleware and other applications on which the program was designed to be used;

• the memory and disk requirements of the program;

• the characteristics of recommended peripherals and the Internet connection.

The wording supplied by the program itself or the related documentation should be used. To reduce ambiguity, the words “system requirements” or an equivalent phrase precede this information. A URI to a Web page giving additional information may be provided.

System requirements are technology dependent and therefore become outdated quickly (see ). But even when the described system has become outdated, information about it may still be useful because the description will enable potential users of the cited information resource to have a general idea of emulation requirements (and choose migration strategy, if emulation is too complicated).

==== Restoration

If the cited information resource is a restored version of the original such as digitally remastered version of a film or a recording, restoration year should also be provided. If the resource has been significantly altered during restoration, or if it has been restored multiple times, restoration year is mandatory.

==== Other information

Other information relevant to the purposes of the particular reference may be included. Such information may include, for instance, links to the datasets the cited information resource is based on, or links to the documentation which enables the usage of the resource.

== Resource types

=== General

• Source: 5.1

• Serialisation: bibitem@type, BibItemType

While does not prescribe a list of possible bibliographic types, there are categories specific to particular information resource categories (5), as well as rendering rules specific to information resource categories (e.g. 5.5: series title appears before item title for broadcasts). For that reason, bibliographic items should indicate what bibliographic type they belong to.

The list of bibliographic types is a union of the types listed in ISO 690 and BibTeX. Master’s Thesis and PhD Thesis from BibTeX are subsumed under “thesis”.

The list maintains a distinction between the following resources, all of which are necessarily included in another resource:

“inbook”

A typically untitled part of a book. May be a chapter (or section, etc.) and/or a range of pages.

“incollection”

A part of a book having its own title.

“inproceedings”

An article in a conference proceedings.

“article”

An article from a journal or magazine.

The above distinction aligns with BibTeX usage.

The overloaded type “electronic resource” should be avoided where possible, particularly if the resource corresponds closely to a paper bibliographic type, such as “book” or “article”. The following types have been added, as specialisations of “electronic resource” with no print or physical counterpart:

• dataset: an electronic dataset, a collection of data that is not meant to be human-readable, and which is typically only machine readable.

• webresource: a human-readable or consumable web resource, a single item accessible online via a web browser, which is not composed of components all of which can be accessed in the same way. Includes media files, as well as individual web pages.

• website: a collection of web resources, presented under the same URL domain and by the same (individual or corporate authorship).

• social_media: one or more resources within a collection that is typically collectively authored by member users. Includes blog posts, forum posts, social media posts, tweets. Is usually a webresource, but not always (e.g. blogs on WeChat are only accessible within the WeChat app.)

• software: computer executable code, not itself human-readable text (though it may generate such text).

The following types have been added, as specialisations of “electronic resource” with a physical but not a print counterpart:

• alert: a single communication intended for multiple persons, and publicly accessible. May be electronic (e.g. Facebook status update) or voice (e.g. evacuation alert), but is typically not print.

• message: a single communication intended for a restricted number of authorised persons (typically one). May be electronic (e.g. Twitter direct message, email) or voice (e.g. a remark made to someone, typically cited as “personal communication”).

• conversation: an exchange of messages between two or more persons. May be electronic (e.g. web chat) or voice (e.g. phone call).

Bibliographic types can overlap; for example much social media can be treated as webresources, and webresources are electronic resources; a Facebook status update can be treated as an alert, or as social media. However each bibliographic type is associated with a particular set of conventions around citation, so classifying an item as belonging to a given bibliographic type determines how it will be cited. Following the classification of citations in , social_media takes priority over other types where it is applicable, particularly for publicly visible communications (such as status updates).

=== Monographs

==== Metadata elements

shows the elements, in order, to be included in references to books and other separately issued monographs.

==== Simple monographs

A printed monograph, such as a book, may be referenced using typical monographic citation rules.

A digital copy of a monograph should be referred to with an online link, if the book is available in the Internet.

==== Composed of multiple volumes

Monographs that are composed of multiple volumes may be indicated so.

==== E-books

Edition, carrier type, media type and/or file format of the cited e-book should be specified, but in a form that does not become obsolete quickly. For instance, giving the precise version of the e-book format or reading device used should be avoided unless it is relevant from the usage point of view.

If the cited e-book does not have page numbers, in-text citation may use any of the following:

• a paragraph number from the beginning of the book, if provided

• a paragraph number within a chapter

• a heading or an abbreviated version of it in quotation marks

If the cited e-book has numbered chapters, sections, and paragraphs, they may be used.

==== Audiobooks

The narrator of the cited information resource should be included if mentioned in the publication. Duration of the audiobook may be specified either in ISO 8601 format or in a more human readable form.

In-text citation for an audiobook should specify the relevant section.

=== Monograph components

==== Metadata elements

shows the metadata elements, in order, to be included in references to component parts within monographs.

==== Component parts of a monograph

If a component part of a monograph (e.g., an article within an encyclopaedia) is cited, information about the host information resource shall be provided along with locality information on the location of the cited information resource within the host information resource.

If the identifier and creator of the host information resource are known, they shall be provided to facilitate identification of the resource.

The locality of the cited component part of an information resource should be stated according to hierarchical navigation, in order to facilitate the reader to locate the cited component part.

Component parts of monographs presented on non-printed medium may be referred to in the same way.

=== Serials and their component parts

==== General

Continuing resources include serials and ongoing integrating resources such as web sites. Examples of their component parts include journal articles and web pages.

Serials have traditionally been published in successive issues which had numerical and/or chronological designations. Publication patterns of digital journals may differ from this; for instance, there may still be volumes and articles but not issues. Citation guidelines must take this diversity into account.

==== Metadata elements

===== Referencing an entire serial or entire issue of a serial

shows the metadata elements, in order, to be included in references to serials.

===== Referencing a component part within a serial

shows the elements, in order, to be included in references to articles or other component parts within a serial.

==== Printed serials

A printed serial may be referred to by the title. The key title and / or ISSN should be used when the title proper is not unique. Other information (publisher, dates of publication) may be provided if appropriate.

==== Non-printed serials

Online serial citations shall contain online access information. If the key title indicates that the resource is available on line, there is no need to repeat this information.

Untraditional numbering schemes of online serials should be recorded as such. Metadata provided in the article (e.g., the publication date and license) may be used in the citation as deemed appropriate.

==== Monograph series

If a component part of monograph series (such as an article or a book) is cited, the series title and number shall be included when present. The monograph series title shall be placed right after the monograph title, indicating that the monograph belongs to a monograph series. Context delimiters such as parentheses should be used to separate the series’ title and the monograph’s title.

A report number, if available, should also be included.

For demonstration purposes, in accordance with , the following examples are shown without the use of the volume or pagination labels.

A serial component part may be available not only as the final published version, but also in unfinished forms, such as preprint.

Since the content and layout of the published resource and its unfinished version may differ, information with which it is possible to identify the edition used shall be included in the citation.

If a component part of an online serial is referred to, PID-based link should be used whenever possible. URI may be used if the resource does not have a persistent identifier, but a link should be provided to an archive copy of the resource as well.

==== Periodical articles

For newspaper articles, publishing date should replace volume and issue.

Digital surrogate of a printed article should be cited as an online resource.

Magazines should be cited either by date or by volume and issue, depending on the publishing pattern.

Newsletters should also be treated like magazines.

=== Programs and applications

==== NEW

• Source: 5.3

System requirements are modelled as a note, prefixed with “System requirements:”. (Note that detailed system requirements are less commonly expressed now for software than they were in the 1990s and 2000s.)

URLs should not be broken at the end of a line: that is a rendering requirement, not an information modelling requirement.

As discussed in , where possible, a more specific category should be used than “electronic resource”, which is properly a description of a medium and not a resource type; monographs and journals are now routinely published online.

The specific type of electronic resource, if not covered by bibitem/@type (e.g. “book”, “dataset”), should be given as medium/genre.

==== General

Programs and applications should be referenced like traditional information resources, such as printed and electronic publications.

Games are considered applications.

==== Metadata elements

Due to the diversity of programs and applications, it is impossible to provide a full list of metadata elements which should or may be used when citing these resources, but for instance, the following elements in may be present:

==== Data elements and guidance specific to type

===== Creator

The name(s) of the creator(s) should be given in preference to that of the distributor, even if the latter is more prominently displayed.

Often a computer program is treated as corporately authored, and neither the designer nor the programmers are credited. In that case the organization publishing the program is credited as the creator.

The system designer of a program should be treated as the creator; other developers involved that are credited in the software are treated as subsidiary creators. If there is no system designer distinct from programmers, the lead developer should be treated as the creator, if present, since they have assumed design responsibility for the developer team.

===== Date of publication

The publication date should be the date (or year) on which the cited version of the program was released for use.

===== Edition and version

Citation should enable identification of the edition and/or version of the program used in order to support reproducibility.

If there is no actionable and persistent identifier,

There is no standard identifier for programs and applications. But Software Heritage Archive ( https://www.softwareheritage.org/), which archives open source programs from various sources, assigns identifiers for information resources harvested into the archive. See https://docs.softwareheritage.org/devel/swh-model/persistent-identifiers.html.

If the citation is not version specific, the publication date may be replaced by the date of citation, which should be the date on which the program was accessed in the context of the material being referenced.

===== License

The license should be described if it is known (see ).

===== Additional information

Reproducibility of results is important. Therefore, citation of a program or application should provide sufficient technical metadata or a link to such metadata, in order to enable access to both the program itself and to its associated metadata, needed for using the referenced program. Such metadata is often available via the landing page of the resource.

If the landing page does not exist or it does not provide sufficient information about, e.g., configuration of the program or acceptable software and hardware platforms, this information should be included either in the citation or elsewhere in the citing publication.

See for additional information.

==== Operating systems

When operating systems are cited, version or edition should be included only if necessary, since such information becomes outdated quickly.

Hardware requirements should be included if the edition / version information is provided.

==== General programs

When programs are cited, version or edition should be included only if necessary, since such information becomes outdated quickly.

System requirements should be included if edition / version information is provided.

==== Games

Electronic games should be referenced using the criteria for programs.

=== Cartographic material

==== General

Examples of cartographic material include maps, atlases, cartographic archival resources and datasets.

==== FROM 6900

• Source: 5.4

The projection, prime meridian, orientation, and reference systems (such as grids and navigational lattices) are given as notes. The spectral characteristics of the sensor, details of the satellite systems, and the amount of cloud cover are also given as notes. Details of the creator’s role are given in bibitem/contributor/role.

==== Metadata elements

shows the elements, in order, to be included in references to cartographic information resources.

==== Data elements and guidance specific to type

===== Title

The form of title used for cartographic material should be one of the following, listed in order of preference (in preference to , ):

1. an overprinted title;

2. a title in the title panel;

3. a title in embedded metadata;

4. a title within the grid or other border which encloses the detail of a map; and

5. a title elsewhere; including external metadata sources.

If there is no clear title in the cited cartographic resource (e.g., atlas or single-sheet map), in addition to the possible titles provided in , a title describing the area covered by the resource may be chosen.

For a map series, the scale and series designation should be supplied, if they are necessary to distinguish the series from others of the same area.

===== Subject and area covered

For cartographic material, the special subject and area covered should be supplied if not given in the title.

The supplied name of a geographical area should be the name of the area as conceived of by the cartographer, whether in their contemporary frame of reference or in a historical sense, rather than as a present-day equivalent.

===== Roles

For cartographic material, the creator’s role may be indicated on the map by terms such as “by”, “cartographer”, “compiled”, “corrected”, “dedicated”, “edited”, “designed”, “drawn”, “copied”, “made”, “engraved”, “revised”, or “updated”.

Attributions may be given in the creator element or in the additional information (see ). The name of an individual who is known to be the major contributor to the creation of a map should be included, even if the name is not prominently displayed.

If both printer and publisher are shown on a map, only the publisher’s name should be given.

===== Series designation

The map series designation, if any, should be given in the form in which it appears in the cited information resource, preceded, if appropriate, by the term “series”.

===== Scale

The scale should be given, if applicable, normally as a ratio (calculated if necessary), e.g., “1:50,000”, or, if given otherwise on the map, it may be cited as it appears. “Scale not shown” or “scale indeterminable” may be used if the scale is missing and it cannot be determined.

If the scale appears in graphic terms, e.g., a scale bar, it may be calculated as a ratio, e.g., “1:63,360” or given in the form on the map, e.g., “scale of miles, 10 [= 60 mm]”.

===== Size

The size of some cartographic resources including printed maps, charts and plans should be given as the height followed by the width, in millimetres, measured between the neat lines. Other dimensions, such as sheet size, may be given, if necessary.

For three-dimensional items, height, width and depth should be given.

For circular maps and globes, the diameter should be given.

===== Spectral characteristics

For remotely sensed cartographic images, the spectral characteristics of the sensor should be given, together with details of the satellite system, if not given elsewhere in the citation. Information about the amount of cloud cover should be given, if appropriate.

===== Additional information

The projection, prime meridian, orientation and reference systems such as grids and navigational lattices may be given if considered important. Sometimes it may be necessary to specify terrain (relief) and colours to distinguish similar cartographic resources from one another.

==== Separately issued cartographic resources

Separately issued cartographic resources may be referred to using the data elements described in this document.

==== As a component part

If a cartographic resource has been issued as a component part within another information resource, it should be cited as part of that information resource. This also applies to the case where the cartographic resource is removed or reproduced from another information resource.

==== Electronic cartographic resources

If the cited cartographic resource is available in the Internet, its URI should be provided.

References of digital surrogates should contain the relevant metadata elements pertaining to the original printed version.

=== Audio-visual material

==== NEW

=== Films, videos and broadcasts

• Source: 5.5

If a single contributor is to be singled out as the creator of a resource, they are assigned the bibitem/contributor/role@type of “author”. Otherwise, if the resource is treated as a collaborative effort, the various contributors should be listed with bibitem/contributor/role@type of “performer”, with their contributions differentiated in the bibitem/contributor/role.

Programmes within a series have the series modelled as a bibitem/series. The title of the series is modelled as bibitem/series/title, and the title of the episode as bibitem/title. The transmitting organisation and channel are modelled as a contributor of type “distributor”.

If an audiovisual resource is a contained item, the bibitem/extent gives its location within the host item; this is typically a timestamp for visual material, and a timestamp or a track for audio material.

==== General

Audio-visual materials include films, videos, radio and TV programmes and broadcasts. For audiobooks, see .

==== Metadata elements

Due to the diversity of audio-visual materials, it is impossible to provide a full list of metadata elements these citations may contain.

shows the elements, in order, to be included in references to audio-visual material.

==== Data elements and guidance specific to type

===== Persistent identifiers

Actionable persistent identifier should be provided as HTTP URI if the cited information resource has a PID.

URI fragment may be used to specify and access a segment within an audio-visual resource such as a video. Such a segment may be for instance a conference presentation embedded in the broadcast covering the entire conference.

===== Roles

Many audio-visual materials are the co-operative output of many individuals, no one of whom has a dominant role as the creator. These resources should be cited with the title as the first element in the reference. If there is a principal creator, such as the director, the resource should be cited with his or her name as the first element.

If one individual is clearly responsible for the intellectual content of an item, he or she should be given as creator in the first element.

Principal performers may be included as well.

===== Date of transmission

In addition to , for audio-visual materials cited from a transmission rather than from the physical item, the date and time of transmission and the transmitting organization should be given, if no other unique identifier of the particular edition is available.

Broadcast date and time should be omitted when citing dynamic sites.

===== Medium, carrier and playback requirements

Sufficient information about physical medium and/or file formats of the cited audio-visual resource should be given if such information is needed for identification of suitable playback devices.

See for a more detailed description of system requirements.

==== Films

For films, the primary date is the year of the initial release in the country of production.

Medium and carrier of the version or edition cited, and its publication date should be included, if features of that version/edition are relevant (if, for instance, commentary only available on the 4K version of a movie is discussed).

Publisher should be included, with the location if relevant and known. If the publisher has changed since the film was originally released and the current publisher is shown, the date should reflect the change.

Local distributor may be indicated in addition to the publisher.

If a digitally restored version of the film is cited, the year of restoration should be provided alongside the year of original release. Restoration year is mandatory if the resource has been restored more than once or if the film has been modified (e.g. colorized) during restoration.

Film includes film of musical performances. Such references should indicate the attributes of both films and musical performances.

==== Broadcasts

Broadcast is a continuous programme stream on certain channel or distribution networks, and it may or may not contain unique programmes with a title that could be referenced in addition to a time range reference of the broadcast segment in question.

Time range specification of a specific programme within broadcast, if included, may be encoded as the URI fragment part of the persistent identifier or URL in the reference. Start time and duration of the programme item may be provided as well.

==== Individual items

Individual items within a programme or broadcast should be cited as component parts, which may contain starting and ending time relative to the start of the programme or in case of a broadcast to absolute time. Duration may be specified after the starting time.

The SMPTE timecode

https://en.wikipedia.org/wiki/SMPTE_timecode

==== Episode in series

For an individual episode within a series broadcasted in TV, the number and title of the season and episode should be given. They should be accompanied with the series title, the transmitting organization and channel, and the full date (i.e. year, month, day and, if necessary, the time of transmission) as expressed in the cited resource and/or in format.

When a series is released on a physical carrier, the publisher and the carrier type should be included.

=== Art and graphic material

==== NEW

=== Graphic Works

• Source: 5.6

The fuller description of a contributors role under bibitem/contributor/role allows for Latin role descriptions such as pinxit and sculpsit.

==== General

Graphic materials are documents in the form of pictures, photographs, drawings, watercolours, prints, and other forms of two-dimensional pictorial representations.

These guidelines apply in citing all works of art in addition to graphic material, such as sculptures.

==== Metadata elements

shows the elements, in order, to be included in references to art and graphic material.

==== Data elements and guidance specific to type

===== Title

The form of title used for graphic material should be one of the following, listed in order of preference (in preference to , ):

1. a title forming part of the original design;

2. a handwritten title on the cited information resource, or custodian’s description;

3. a popular or traditional title, e.g., Mona Lisa;

4. a title in embedded metadata; and

5. a title elsewhere; including external metadata sources.

If there is no clear title, in addition to the possible titles provided in , the title can describe the object depicted (e.g., the person, building, location).

The subject matter and print process should also be supplied, if available.

===== Roles

Several individuals may be concerned in the production of graphic material. If their names are recorded on the item, they should be included with an indication of their roles.

Abbreviations of roles, such as for print addresses, are allowed. For instance:

• The person who printed the lithograph may be indicated by abbreviations such as “lith.”, “lith. impr.”, “litho.”, “lithog.”.

• The person who created the original art may be indicated by abbreviations such as “invt.” or “inv.” (for “invenit”).

• The person who commissioned the work may be indicated by abbreviations such as “f.f.” (for “fieri fecit”).

===== Location

If the location of a graphic work in a museum or repository is essential to its identification, physical location should be included.

===== Size

The size of a graphic work may be specified as the height followed by the width, in millimetres, measured between the neat lines. Other dimensions, such as sheet size, may be given, if necessary.

===== Catalogue and catalogue number

Collected graphic and art material are often listed in catalogues. The catalogue may be specified as additional information after the title, prefixed with the phrase “Catalogued in:” or its equivalent.

Since there is a relation between the cited graphic work and the catalogue or catalogues in which the work is described, rules in apply.

The catalogue number, called the “catalogue raisonné” (“reasoned catalogue” in English), is used to identify a specific item within a catalogue. It should be given after the catalogue in order to facilitate location.

==== Individual works

A graphic or artwork may be individually referenced using data elements described in this document.

Digital surrogate information can also be provided as online access locations.

==== As a component part

If a graphic work has been removed or reproduced from a known book or other information resource, it should be cited as part of that information resource.

If the component part and the host resource have the same creator, it is not necessary to repeat the name.

==== Temporary or destroyed work

It is possible to cite surrogates of installations and other works of art which were temporary (or have been destroyed) and no longer exist.

=== Music material

==== NEW

=== Music

• Source: 5.7

• Serialisation: bibitem/contributor/role

Librettists and composers are both contributors of role type “author”; the role of composer and librettist needs to be differentiated in the contributor role for proper rendering.

The type “music” covers both music scores and music recordings, as is done in . The two are differentiated in medium; following the practice in , medium/content “notated music” is used for music scores, while medium/carrier is used for recordings (audio disc = CD, online resource). While it is possible to cite music in the abstract, without referencing either a score or a recording (especially if the score is not published), this is not good practice, for the same reason that particular editions are required for citation of textual works.

==== General

Citable music material includes printed resources (sheet music), recordings on hand-held devices and on-line resources.

==== Metadata elements

shows the elements, in order, to be included in references to music material.

==== Data elements and guidance specific to type

===== Title

The title for classical musical works must include the opus number or catalogue number where available.

===== Creator

In references to written music, the composer should normally be shown as the creator, and the librettist or lyricist as the (first) subsidiary creator, even if a librettist or lyricist is given equal prominence in the source.

==== Performed and recorded music

In references to recorded music, the composer should be shown as the creator, and the librettist or lyricist, as in , should appear as the first subsidiary creator, if applicable. The conductor, soloists and other artists or groups involved with the recording should be named as subsidiary creators. Creator and contributor information should enable unique identification of the information resource; this may require even including persons in other roles, such as cantors.

==== Printed music

For printed music, a score that is musically identical with a full score, but reduced in size, may be described as a miniature score or pocket score if it does not exceed 200 mm in height, or as a study score if it exceeds 200 mm in height.

If a printed score has an identifier, it should be included. The identifier is usually an ISMN [International Standard Music Number], but a printed score may also have an ISBN in addition to or even instead of ISMN. If ISMN is available, it should be used instead of the ISBN.

Multiple creators of a musical score may be indicated, with their roles.

Digital surrogates of music scores may be provided via specification of the online access location.

==== As a component part

If the cited resource is a part of a larger work, or one of the works published together in an aggregation, information about the musical work or recording should be included in the reference. For example, a track within an album. The creator name may be repeated if deemed necessary.

In sheet music, if reference to a particular bar is required, the bar count before or after rehearsal marks should be used. Bar numbers should not be cited unless they are printed on the score.

=== Patents

==== NEW

=== Patents

• Source: 5.8

Patents are encoded similarly to contributions within a serial (5.2), with the series designation for the patent preceded by the country name or code. The date of application is encoded as a date of type “issued”.

==== General

There are two kinds of patent publications: patent applications and issued patents. In some countries only patent applications are published.

==== Metadata elements

shows the elements, in order, to be included in references to patents.

The name of the proprietor or applicant creator (see ) should precede the title of the patent document. The country name or code (see ) and the official designation of the series within which the patent document is numbered (host document) should follow. The patent serial number should suffice as numeration. The patent number should be presented as in the original patent application / publication.

The order of elements in a reference to a patent is, with the exception of the date, the same as for a reference to a component part within a serial (see ).

==== Data elements and guidance specific to type

===== Type

If it is not already clear in the reference, the fact that an item is a patent should be stated.

===== Parent body

For patents, the country code or the World Intellectual Property Organization (WIPO) ST3 code may be used to replace the name of the country of origin or originating office.

===== Patent inventors

Inventors of the patent may be shown in the reference as subsidiary creators with the role “inventor”. This is unnecessary if the patent application was filed by the same individual.

===== Additional information

Additional information may be given between the title and the official designation, such as:

• any classification symbols, and

• the dates of application and/or issuance.

===== Patent number

Relevant patent numbers should be provided in the reference, such as the patent application number, and the patent issuance number.

==== Patent applications

The date of patent application submission may be provided preceded by “Application date:” or its abbreviation “Appl:”.

==== Issued patents

The date of patent issuance may be provided preceded by “Issuance date:” or its abbreviation “Iss:”.

=== Reports in series and similar information resources

==== General

Reports and similar information resources issued in series by research bodies, standards bodies and similar organizations are often identified by the series title and number. In references to such resources, the series title and number are essential elements.

==== Metadata elements

shows the elements, in order, to be included in references to reports.

==== Data elements and guidance specific to type

===== Series title

Series title may be abbreviated in a generally recognized form, e.g., “RFC” for the “Internet Engineering Task Force’s Request for Comment series”. Additional information may be provided as appropriate; for instance, there are several categories of RFC’s, and citations may specify the category.

===== Edition

If a specific edition and/or part of the information resource is cited, the edition and/or part should be indicated as in the original resource (with e.g., the version number or year).

If the edition is not given, the citation refers to the latest edition of the information resource.

==== Standards

===== General

The name of the standardisation organization should be provided as specified in .

Standards may be cited in an undated reference, by omitting the edition in the citation. An undated reference covers the latest edition with published amendments and corrigenda that apply. However, if there has been a major change with the latest edition of the standard, such as if the standard is split into two or more parts, the citation no longer covers that edition.

A dated reference of a standard specifies a particular edition. Such reference only covers the stated edition; neither future modifications of that edition (corrigenda and amendments) nor future editions of the standard will be included.

In a dated reference, all published corrigenda and amendments may be included by adding a phrase equivalent to “(inclusive)” after the edition, in the language of citation.

If a standard has more than one active identifier from the standardisation organization, both identifiers should be shown

===== Jointly published standards

If the standard is a joint publication by two or more standards bodies, one of them may be selected as the primary publisher.

A joint publication may be assigned different publication identifiers by different standardisation organizations. In this case, these identifiers should be displayed for clarity, separated by a “|” symbol.

===== Dual-published standards

If the same standard has been published by two or more standards bodies, the standard used should be cited, except if it has access limitations and the other standard does not. In that case, both versions may be listed.

===== Draft standards

Citation guidelines published by the standards body shall be followed. For instance, according to the ISO guidelines on citing standards and their draft editions (see ), citing International Standards, Final Draft International Standards (ISO/FDIS) and Draft International Standards (ISO/DIS) is allowed, but earlier versions shall not be cited.

An ISO/DIS or even an ISO/FDIS specification may still change significantly before it is published as an International Standard. Therefore it is recommended to only refer to final published ISO deliverables, such as International Standards, Technical Specifications and Technical Reports.

It is acceptable to cite ISO and IEC deliverables under preparation, according to ISO/IEC Directives Part 2, once they have reached the enquiry or approval stages of the International Harmonized Stage Codes. DIS (Draft International Standard) stage. WDs (Working Drafts) and CDs (Committee Drafts) shall not be cited.

Once an international standard has been published, its DIS and FDIS versions published during the preparation of the standard shall no longer be cited.

ITU draft Recommendations may be cited as “Draft new Recommendation” with its provisional name.

IETF (Internet Engineering Task Force) Internet-Drafts (working documents) may be cited as “work in progress”. The edition of the draft shall be included after the title as additional information. If the draft is an individual contribution, the author or authors shall be the primary creators.

If a national or international standards body does not have published guidelines on how to cite their draft specifications, maturity level of each draft has to be estimated separately, on the basis of e.g. voting results and the number and nature of comments. If there is a possibility that the standard will not be approved, its draft versions should only be cited if necessary.

===== Adopted standards

National standardization bodies publish adoptions or translations of International Standards as national standards. These specifications can differ slightly from the original document.

If an adopted or translated International Standard has been referenced, the national standard shall be cited instead of the original International Standard.

===== Multi-part standards and standard series

Some standards are published in multiple parts, forming a standard series. Each of these parts are potentially published at different times. A citation to the standard series itself is considered to include all amendments and corrigenda of all cited parts.

It is possible to cite multiple parts of a standard series using an explicit reference. These parts may be indicated in the standard series’ document identifier, and all part titles should be omitted from the citation.

In cases where part numbers are not contiguous, or where a citation wishes to cite multiple parts but not all parts of a standard series, the parts cited may be explicitly described in the document identifier.

It is possible to make a blanket reference which covers all parts of a standard series. Such reference should be indicated by a phrase equivalent to “(all parts)”, in the language of citation, placed after the standard series’ document identifier, while omitting all part numbers and titles.

Some multi-part standards do not have consistent document identifiers. For instance, ISO 19157:2013 and ISO 19157-2:2016 both belong to the same standards series, but the former does not have a part number of 1, as it was originally intended to be a standalone standard until part 2 was developed.

When the document identifiers differ within the same standard series, the standards should be cited explicitly to maximize clarity. The “(all parts)” syntax should not be used to refer to the standard series.

An undated reference with a standard series that contain different identifiers only covers the last edition which was not yet published in two parts.

Some standardization organizations organize their standards documents in series and sub-series. References to them can be made by utilizing the standard series (or sub-series) identifier.

=== Archival materials

==== NEW

=== Archival materials

• Source: 5.10

The fonds or collection of an archival item is modelled as a series. The box and folder number is encoded as an access location. The reference code is encoded as an identifier. The archive is encoded as a contributor of type “distributor” (since the archive makes the resource available, even if to a restricted public).

==== General

Archival materials may be individual documents or collections of documents. They may be printed or born digital resources, and there are also digital surrogates of other types of material, such as manuscripts or art.

An archival information resource may be a publication; e.g., a book, article or graphical work, and if so, it is referenced accordingly. Normally this is not the case and archival resources usually require additional information (such as fonds/collection or geographic location).

Archival materials include all kinds of governmental publications but also private documents such as letters, diaries or other textual documents; photographs, and various (other) ephemeral resources, or digital surrogates of these resources.

This document does not cover citing ephemeral resources outside the archival context.

==== Metadata elements

shows the elements, in order, to be included in references to archival information sources.

==== Data elements and guidance specific to type

===== Title

Titles of archival materials are sometimes vague (e.g., “Correspondence”) and therefore additional elements should be included in the citation whenever necessary.

===== Creator

Personal and / or organizational authors should be included when possible.

If the host organization does not own the cited information resource, the owner should be mentioned (if known and if the information is not confidential).

==== Individual documents

If a hand-held document is cited, the location information has to be sufficiently precise.

For digital information resources, URI may replace the physical location information.

If the cited digital document is a surrogate, the location of the original document may also be given.

References to born-digital documents may be made using data elements specified in this document, including digital identification mechanisms.

==== Collections

Collection level citations should specify the creator and the title of the collection and the location.

If the cited collection is part of a larger collection, both collections should be specified.

=== Research datasets

==== NEW

=== Datasets

• Source: 5.11

The organisation publishing or producing the dataset is modelled as a publisher; the organisation hosting the dataset in a repository is modelled as a distributor. Data compilers, editors, investigators, and data managers are all modelled as editors. Timestamps are modelled as extents; queries wherever possible should be preented as URIs. Where that is not possible, the description of the necessary query to access the required data subset should be presented as an accessLocation.

==== General

Research datasets are usually electronic information resources, but they may be hybrid or even traditional (printed, hand-held). Whatever the form, traditionally datasets, like computer programs, have not been cited in publications. However, they should be treated in the same way as conventional publications, in order to give due credit to creators of these materials.

This document provides guidelines only for citing electronic research datasets; more information about citing other kind of datasets are available from other sources.

Since datasets are very diverse, no guidelines are provided on how to cite particular categories of data, with one exception (see on citing research interviews).

In addition to the actual data (possibly in different versions and in multiple file formats) datasets often contain related items, such as codebooks and (in e.g., social sciences) questionnaires used for data collection. Programs and documentation for data analysis may also be available. These related items are often linked to a landing page to which the PID resolves. All these items may be cited independently.

Other dataset citation guidelines include Digital Curation Centre’s guide “How to Cite Datasets and Link to Publications” ; and Data citation principles published by FORCE11

https://www.force11.org/group/joint-declaration-data-citation-principles-final

==== Metadata elements

shows the elements, in order, to be included in references to research datasets.

==== Data elements and guidance specific to type

===== License

The licenses of both the data itself and related items such programs should be included in the reference if these resources are not in the public domain.

===== Roles

People responsible for the cited dataset and related items may include for instance investigators, technicians, data managers, compilers, editors and programmers. Names and roles of creators mentioned in the research dataset documentation and the full name of the project or organization responsible should be provided.

A name identifier (e.g. ISNI, ORCID) should be provided in brackets after the name if known. If the host organization changes its name or is merged with another organization, the identifier makes it possible to find the name of the current organization.

The name of the host organization or organizations shall be given. Identifier, if any, should be provided.

If the cited research dataset has been created by a project and no personal authors are described, the project may be specified as the author. Since projects are usually short-lived, the name of the organization hosting the project shall then be provided as well.

Publisher or distributor of the dataset, such as a data repository, shall be included in the reference if it is different from the host organization.

===== Persistent identifiers

Persistent identifiers such as DOI names shall be given, if available. If the dataset has also a traditional standard identifier, it should be included as well.

===== Linked resources

A user may need to access various other resources before it is possible to utilize the dataset or the tools for processing the data. These resources include user guides, codebooks and questionnaires.

Software user guide citations shall contain a link to an archived copy, since these documents are short-lived; when the application is updated, the user guide changes as well.

If the cited dataset is part of a series, it should be given.

===== File size

File size of the dataset shall be included in the reference after the URI if the dataset is “big data”,

https://en.wikipedia.org/wiki/Big_data

===== Dataset and publications

It may be necessary to cite both the publication based on the data, and the research dataset. These interlinked citations establish a connection between the data and the results based on the data.

===== Provenance information

Research data provenance is information about the inputs, entities, systems, and processes that influenced the cited dataset. Provenance information assists reproduction of the data and therefore, has a major role in increasing trust and credibility. It provides also a historical record of the data and its origins.

Provenance information may be provided as plain text or as structured metadata, based on for instance W3C PROV data model.

https://www.w3.org/TR/prov-dm/

===== Edition and version

Research datasets are often dynamic: new data is constantly added and past errors corrected. A single dataset can also be available in different versions, enabling variety of uses, and in multiple file formats. If there are several options, data used shall be cited in such a way that it is possible to identify the version and / or file format used from the reference.

If only a subset of the dataset has been used, this subset should be specified. Following the Research Data Alliance Working Group on Data Citation recommendation

https://rd-alliance.org/system/files/documents/RDA-DC-Recommendations_151020.pdf

Citations like the one above should facilitate re-creation of the dataset cited, even when this requires technological changes such as migration.

This is possible if datasets are time-stamped and versioned, and if there are time-stamped queries with which to retrieve the desired subset of the data. This approach is promoted by the Research Data Alliance Data Citation WG

=== Web sites and their component parts

==== General

In this document, Web resource is any asset that can be identified, addressed or handled in some way on the Web or in a Web archive. Such resources are often dynamic, which makes citing them a challenge. For instance, newspaper home pages can change several times per day. There are of course also static documents; for instance, doctoral dissertations never change even on the Web, although their locations can.

==== Metadata elements

shows the elements, in order, to be included in references to web sites and web pages.

==== Data elements and guidance specific to type

===== Archive reference

A Web archive reference should provide the name of the Web archive as the distributor, the date and time

If a Web page consists of several components, different files may have been harvested at different times. Due to this, archived copy of a page may not be fully synchronized.

===== Edition or version

Reference to a past edition or version of a Web information resource should specify edition or version (e.g. by number) and/or date and time in an appropriate manner and provide a link to the archived copy used.

Web archives, such as the Internet Archive (http://archive.org) allow the users to search archived copies of documents with their URIs. Users may also make archival copies of web pages. If the link in the reference is dead or if the retrieved document is no longer the cited document due to content drift, it may still be possible to retrieve the cited document from an archive. Referencing archived Web documents increases precision and persistence, but the latter depends on sustainability of the Web archives.

==== Web site

If the reference covers the entire web site, a coverage note should be supplied after the web site title.

If there is any doubt that both the cited information resource and its URL will be stable, citation should include a link to an archived copy of the resource. The original location may be provided as well.

In an inline reference, if the context makes it clear that the entire site is being referenced, it is sufficient to give just the name of the service and the URI of the entry page.

If the cited information resource is a defunct web site, such as the domain having been taken over by another host, the time span of the relevant archived copies should be specified.

==== Web archives

Time span of archived copies may be provided.

See for more information on the use of Web archives.

==== Web page

The citation shall contain the name of the page / site and the dates it has been used and archived.

If the cited Web page is dynamic and a specific version is used, an archived version of the resource shall be cited. If an accurate archival copy of the page is not found, one shall be made.

If version information is not relevant, archive reference may be omitted.

If all versions of a dynamic pages are cited simultaneously, Web archive citation shall be generic and specify the date range covered (see and ).

There are Web sites (belonging for instance to newspapers) that are harvested several times per day. Web archives may have thousands of copies of e.g., home pages of these sites. Almost all of these copies will be different within one archive, but due to harvesting techniques, even the copies in different archives harvested at (almost) the same time may differ from one another. Only the archival copy actually used should be cited, not a copy from another archive even if it has been harvested at about the same time. If for any reason it is necessary to cite another archived copy, it is important to check that it contains the cited information.

Web archives usually do not check if the harvested Web page has changed since it was harvested last time. Some or even all harvested copies of the page can be identical. Even in such cases it is important to include the date, since the page may change in the future.

==== Component part of a web page

If it is necessary to specify a component part within a larger resource such as a Web page, its media type may be supplied in brackets, e.g., [still image]. Note that a component part can be a part of a resource which is itself a component part.

=== Social media and services

==== NEW

• Source: 5.12

Social media items are typically items of bibitem/@type “webresource”, although they may not always have an associated exposed URI (particularly for social media accessed primarily via phone app, rather than web browser). As with , a distinction is made between items publicly available (or at least, broadcast to a defined group of people), and one-on-one text messages; the latter are conflated with voice and face-to-face communications between individuals, and are cited as personal communications.

A person’s real name may be related to the online identity assumed by the author of the item, if both are known. The online identity is treated as a pseudonym of the real life identity, and is serialised in the same way ().

The particular form of social media communication (e.g. blog post, forum post, tweet, message) should be given through medium/genre. At a lower priority, the service name or type should be encoded as medium/form, since it relates to how the content is to be accessed by an intermediary app. (For example, a tweet is not just accessed by a web browser, as a web resouurce; it is accessed and rendered by a web browser in the context of Twitter.)

==== General

Information shared via social media and services is frequently used in research, both within and outside the academic community ().

This document does not specify rules on what kind of social media may or may not be cited, but if the information is not publicly available, it should be cited as personal communications. Organizations may specify their own limitations on what they think is acceptable.

Guidelines listed below are partially based on the ones published by the American Psychological Association .

Social media services include Twitter, Facebook, LinkedIn, Instagram, Snapchat, etc.

==== Metadata elements

shows the elements, in order, to be included in references to social media and services.

==== Data elements and guidance specific to type

===== Creator

A personal author’s last name and initials, or organization’s full name should be provided. A name identifier may be provided in brackets after the name.

Following the real name, social media identity information should be supplied when available. The author’s screen name should be given as associated information (in brackets). The screen name is considered to be the user identifier that allows others use to address the creator on the service. It is preferable to prefix the service name to the screen name.

The author’s screen name is provided to facilitate communication to the creator.

The poster is the person who posted the content, not necessarily the person who created it. The real creator should be included in the reference when applicable.

In a citation, if the service name is provided in other parts of the reference, omission of the service name in the screen name is possible.

===== Title

The name of the page or the content or caption of the post should be specified as the title. Translation of the title may be provided after the title.

The titles of status updates, tweets, pages, blog posts or photographs should not be italicized; the titles of items that stand alone, such as videos and photo albums, should be italicized.

If the item contains no words (e.g., a photograph without a caption), a description of the item should be provided as a supplied title.

===== Medium

Medium designation (e.g., Tweet, Facebook status update, photograph, blog post, timeline, video file) should be provided after the title.

===== Service name and type

Service name or type should be specified. This information is especially important if it might otherwise be difficult to tell the nature of the resource.

===== Date

Full date should be specified in ISO 8601 format for items that have a date associated with them; otherwise, the year is sufficient.

The date and time the content has been viewed should be provided in addition to the date associated with the post.

The accurate time should be provided in the format when citing social media.

===== Network location

Network location of the cited resource shall be provided.

When a social media Web site is first mentioned in the text, just the URI may be provided.

===== Archived copy

Like other Internet resources, social media may be short-lived. Therefore, if there is a copy of the resource in a Web archive, URI to the archived copy may be provided as well.

If there is a reason to believe that the resource will suffer from link rot or content drift, only the URI of the archived copy may be given.

Generic Web archives do harvest social media resources, but there is a subscription-based program for archiving social media, ArchiveSocial.

==== Services

Record aggregators, such as blogs, message boards and forums, can all be considered services. They may be cited using data elements provided in this document. Forum posts and comments may be referred to in the same way.

Inline citations may be made by referencing the URIs of the service or of the service accounts.

==== Stream of records

A stream of records may be referenced by referring to their appropriate URIs.

==== Individual records

Individual records may be referenced on a service using data elements provided.

Some records such as posts and status updates do not usually have titles. The message body may be treated as its title; though truncation may be required in citation.

Some social media allow direct links to comments; but some do not. The following example indicates the limitations when there is no direct link and that the comment date is not precisely dated.

Newsgroups should be treated as serials. USENET posts should be cited by date and time.

==== Posting of content belonging to an original creator

The poster and original content creator should be distinguished. The original content and its creator should be cited as part of the information resource being cited.

A relationship binds the cited information resource with the original information resource, therefore rules in apply. Relation labels may be created as appropriate to facilitate citations.

=== Unpublished information resources

==== General

Unpublished documents are a diverse group of information resources, including for instance personal communications, but also medieval or modern manuscripts, raw scientific data, preprints of scientific publications, doctoral dissertations and master’s theses.

==== Metadata elements

shows the elements, in order, to be included in references to unpublished information resources.

==== Data elements and guidance specific to type

===== Resource type and location

Resource type and the location or the name of the service via which the cited resource is available should be included.

Resource type specific metadata elements, such as manuscript shelf mark, should be cited according to the practice of the memory institution which owns the resource.

==== Personal communications

Personal communications include for instance letters, emails and diaries, but also lectures and conversations, either face to face or via telephone.

Unpublished personal communications should be referenced in text only, with the exact date and, if necessary, time in format. A published, archived or recorded information resource (e.g., a letter, diary or performance) should be cited accordingly (e.g., as a publication, archival information resource or as an audio-visual material).

Inline citation examples follow.

==== Group communications

Group communications, such as mailing lists, may be cited in a similar way as with serials.

==== Dissertations and theses

The resource type of doctoral dissertations and master’s theses should be indicated in the citation.

==== Manuscript

Medieval manuscript folio numbers, including the recto/verso reference, should be provided when the citation does not cover the entire manuscript.

If there is a facsimile of the cited medieval manuscript, it may be mentioned in the reference.

==== Preprint

If a preprint is cited, the name of the access provider and the repository should be included in the citation. If there are several versions of the preprint, version number of the cited version should be included as well.

Publication process information should be provided if available. For instance, if a cited version of a scientific article has already passed peer review and has been submitted for publication, a phrase “Manuscript submitted for publication” may be used.

There are different policies regarding the usage of some unpublished resources, such as preprints. Some publishers may forbid their usage, while some others may require that preprints are treated as personal communications (that is, cited in the text only). This document does not provide guidelines in this respect, but it should be remembered that although most preprints will eventually be published, significant changes may take place during the peer review process. For instance, the published version of the article cited in (see https://doi.org/10.1103/PhysRevA.86.031602) is quite different from the preprint.

==== Phone calls

A personal phone call may be referred to, both for a transcript and its recording.

==== Presentation

Presentations may be referred to as unpublished information resources.

A presentation video may also be referred to.

Presentation slides may be cited as well.

== Reference model XML serialization

The following is a sample serialisation of in XML. This is not intended to be the only possible serialisation of , whether in XML or in another format. In fact, existing reference tools such as BibTeX are expected to be used as serialisations. The sample serialisation is intended to provide illustrations of the model, and to discuss the model succinctly by referencing its elements. (Of course, the serialisation can be used in practice as a basis for data exchange of references.)

The serialisation is the canonical grammar of the relaton-models repository, included verbatim below; refer to for an introduction to the format. The repository grammar is the single source of truth for the serialisation.

== Mapping to BibTeX

This listing includes the BibTeX fields presented as part of the LaTeX package, and (asterisked) the additional BibTeX fields in citavi .

=== Categories of information resource

From .

=== Creator

From .

Note that BibTeX processes names as follows:

=== Title

From .

=== Document relations

From .

=== Medium

From .

=== Edition

From .

=== Production

From .

=== Numeration and pagination

From .

=== Series title and number

From .

=== Identifiers

From .

=== Location

From .

=== Additional general information

From .

==== Classification

==== Price and availability

== Bibliography

• [], BibTeX.

• [], Citavi.

• [], IFLA. Functional Requirements for Bibliographic Records. IFLA Series on Bibliographic Control 19. Munich: K.G. Saur. 1998.

• [], Lamport, Leslie. 1986. LATEX. A document preparation system. User’s Guide and Reference Manual. Reading, MA: Addison-Wesley.

• [], OASIS. RELAX NG Compact Syntax. Edited by James Clark. 2002-11-21.

• [], OASIS. RELAX NG Compact Syntax Tutorial. Edited by James Clark, John Cowan, MURATA Makoto. 2003-03-26.

• [], Freedm N., Borenstein, N. Multipurpose Internet Mail Extensions (MIME) Part Two: Media Types. November 1996.

• [], Berners-Lee, T., Fielding, R., Masinter, L. Uniform Resource Identifier (URI): Generic Syntax. January 2005.

• [], Nottingham, M. Web Linking. October 2017.

• [], W3C. XML Schema Part 2: Datatypes Second Edition. W3C Recommendation. 2004-10-28.

• [], URL Shortening. Wikipedia, 2019-08-08. Archived copy available from: Internet Archive [distributor], [viewed 2019-08-13].

• [], PURL, Persistent Uniform Resource Locator [online]. Wikipedia, 2019-08-08. Archived copy available from: Internet Archive [distributor], [viewed 2019-08-13].

• [], Permalink [online]. Wikipedia, 2019-08-12. Archived copy available from: Internet Archive [distributor], [viewed 2019-08-13].

• [], Perma.CC homepage [online]. Available from: [viewed 2019-08-13].

• [], FIPS 202, Federal Information Processing Standard 202, SHA-3 Standard: Permutation-Based Hash and Extendable-Output Functions. Gaithersburg: Department of Commerce, USA. August 2015. Available from:

• [], Social media data research and use. [online]. NCSU Libraries, [s. a]. Available from: [viewed 2019-08-13].

• [], IANA Media Types. Internet Assigned Names Authority. Available from: [viewed 2019-08-13].

• [], PRONOM. National Archives, UK. Available from: . [viewed 2019-08-13].

• [] BALL, A. & M. Duke, (2015). ‘How to Cite Datasets and Link to Publications’. DCC How-to Guides [online]. Edinburgh: Digital Curation Centre. Available from: http://www.dcc.ac.uk/documents/how-guides [viewed 2016-07-08].

• [] LEE, Chelsea: How to Cite Social Media in APA Style (Twitter, Facebook, and Google+ [online]. American Psychological Association, 2013. Available from: http://blog.apastyle.org/apastyle/2013/10/how-to-cite-social-media-in-apa-style.html [viewed 2016-07-07].

• [] Recommended standard on two-letter codes for the representation of states, other entities and intergovernmental organizations. Revision adopted by the SCIT Standards and Documentation Working Group at its eighth session on 22 March 2007. (WIPOStandardST3). In: WIPO Handbook on Industrial Property Information and Documentation [online]. Geneva: World Intellectual Property Organization April 2007. Available from: [viewed 2019-02-07].