Syntax and Vocabulary of the Academic Metadata Format

Tim D. Brody, Zhuoan Jiao, Thomas Krichel and Simeon M. Warner

draft: 2003–01–30

This document is available in PDF format for US letter size paper and for A4 size paper. We are grateful for helpful comments from José Manuel Barrueco Cruz, Christopher F. Baum and Ivan V. Kurmanov.

1: Introduction

This document is a draft for the Academic Metadata Format (AMF). AMF encodes descriptions of

AMF relies as much as possible on standard vocabulary by borrowing metadata terms from other vocabularies. The relevant vocabularies are

AMF is not definite and may change at any moment. The current specification is provided as a basis for experimental deployment only. During this test phase, the draft standard's files are maintained at Work on the AMF is supported by the Open Archives Initiative.

The remainder of the document is organized as follows. In Section 2, we introduce the general markup of AMF. In Section 3, we describe the names and semantics of elements used by AMF. In Section 4, we discuss constraints on the contents of elements. Such value constraints are indicated by the use of italics in the description of element semantics in Section 3. In Section 5, we present optional attributes that may be useful to further qualify element contents. In Section 6 we give some examples.

2: AMF markup

AMF is encoded in XML. All element names used by AMF use lowercase letters only. The AMF data is enclosed in an AMF root element that has the name amf. This element may have three attributes.

  1. xmlns—the value of which must always be the AMF namespace URI:

  2. xmlns:xsi—the value of which must always be the namespace URI for XML schema:

  3. xsi:schemaLocation is a pair.
    The first part is the AMF namespace URI (matching xmlns definition):
    The second part is the URL of the XML schema of the AMF response:

AMF is an open vocabulary in the sense that the AMF XML schema file allows to place elements from foreign vocabularies within the AMF vocabulary. This can be done on the root level of the AMF element, or in any child contents of that root element. Foreign element names must be namespace qualified.


<amf xmlns=""
    <text id="oai:arXiv:hep-lat/0008015"> <title> ... </title> ...

In the example, text is a "noun" element. The AMF root element must contain one or more nouns. Nouns are repeatable. There are four nouns:

person a physical person
organization an entity that has physical persons as its members
text a dctype:text
collection a dctype:collection of resources

Each instance of a noun element in AMF data that is not an empty element is called an AMF record. All child elements of AMF records are optional and repeatable. An AMF record admits two types of child elements.

The first type are "adjective" elements. Adjectives give further information about nouns. Some adjectives have a nested structure. In the example above, title is an adjective.

The second type are "verb" elements. Verbs relate one noun to other nouns. Each verb must have one or more nouns as children. Verbs must not have adjectives as direct children.

3: The elements of AMF

3.1: The person and organization nouns

The person noun element describes or refers to a physical person:

<person id="..."> ...adjectives and verbs... </person> or <person ref="..."/>

The organization noun element describes or refers to an organization. An organization is a group of two or more persons:

<organization id="..."> ...adjectives and verbs... </organization> or <organization ref="..."/>

Both nouns accept the same verbs and adjectives. Therefore they will be collectively referred to as the "p/o" noun in the remainder of this document.

3.1.1: The adjectives of the p/o noun

<name> unstructured full name, as vCard:FN </name>
<shortname> short name e.g. IMF, as vCard:NickName </shortname>
<familyname> full name, as vCard:N;FamilyName </familyname>
<givenname> given name, as vCard:N;GivenName </givenname>
<additionalname> additional name, as vCard:N;AdditionalName </additionalname>
<nameprefix> honorary prefix, as vCard:N;HonoraryPrefix </nameprefix>
<namesuffix> honorary name, as vCard:N;HonorarySuffix </namesuffix>
<date> date associated with the p/o, as dc:date </date>
<homepage> URL of homepage </homepage>
<postal> postal address as vCard:LABEL </postal>
<phone> telephone number as vCard:TEL;TYPE=pref,voice </phone>
<fax> fax number vCard:TEL;TYPE=pref,fax </fax>
<email> email as vCard:EMAIL;TYPE=internet,pref </email>
<identifier> for the p/o from a scheme that does not use AMF, as dc:identifier </identifier>

3.1.2: p/o to p/o verbs

<isreplacedby> an organization is replaced by another </isreplacedby>
<replaces> an organization replaces another </replaces>
<ispartof> an organization is a part of another </ispartof>
<haspart> an organization has another as a part </haspart>

3.1.3: p/o to text verbs

<isauthorof> as dc:creator </isauthorof>
<iseditorof> as dc:creator or dc:contributor </iseditorof>
<ispublisherof> in the sense of dc:publisher </ispublisherof>
<istranslatorof> </istranslatorof>
<ismaintainerof> p/o who maintains metadata about the text </ismaintainerof>

3.1.4: p/o to collection verbs

<iseditorof> p/o responsible for the contents of the collection </iseditorof>
<ispublisherof> in the sense of dc:publisher </ispublisherof>
<ismaintainerof> p/o who maintains metadata about the collection </ismaintainerof>

3.2: The text noun

The text noun element describes or refers to a text, independent of its status. Thus a PhD thesis, an article in a learned journal, the transcript of a speech etc, are all texts:

<text id="..."> ...adjectives and verbs... </text> or <text ref="..."/>

Note: A journal is not a text, it is a collection. A book may also be a collection if it contains papers by different authors.

3.2.1: The adjectives of the text noun

<title> as dc:title </title>
<abstract> as dcq:abstract </abstract>
<keywords> list of uncontrolled keywords, may be subject to a scheme vocabulary to be developed </keywords>
<classification> list of classification codes, see section </classification>
<copyright> a plain-text statement about the copyright, as dc:rights </copyright>
<status> a plain-text description of the status of the text, say published in a journal,presented at a conference etc.</status>
<comment> something about the text that is not the status, e.g. a dedication </comment>
<email> email for the text, not necessarily one of the authors or editors </email>
<date> date associated with the text </date>
<displaypage> URL of a page where access to the text is explained </displaypage>
<citation> unstructured full text of citation </citation>
<serial> container tag for structured serial access information that citation can provide
    <journaltitle> title of serial, as OpenURL:title, dccite:journaltitle </journaltitle>
    <journalabbreviatedtitle> abbreviated title of serial, as OpenURL:stitle, asdcite:journalabbreviatedtitle </journalabbreviatedtitle>
    <journalidentifier>identifier (usually ISSN) of journal,as dccite:journalidentifier </journalidentifier>
    <issuedate> date on the serial issue cover, as OpenURL:date, as dccite:cronology </issuedate>
    <volume> as OpenURL:volume, as dccite:volume </volume>
    <part> as OpenURL:part, as dcite:number </part>
    <issue> as OpenURL:issue, as dcite:number </issue>
    <season> season of publication (spring or summer or autumn or winter),as OpenURL:ssn, as dcite:cronology </season>
    <quarter> quarter of publication (1 or 2 or 3 or 4), as OpenURL:quarter, , as dcite:cronology </quarter>
    <startpage> number of the first page of the text in the serial issue, as OpenURL:spage </startpage>
    <endpage> number of the last page of the text in the serial issue, as OpenURL:epage </endpage>
    <pages> unstructured page data, as OpenURL:pages, as dccite:pagination </pages>
    <articlenumber> article number, in the absence of pagination, as OpenURL:artnum </articlenumber>
<file> a container for full-text file information; it may be repeated for each component file.
    <url> URL for the file itself </url>
    <function> the function of the file in the text, e.g. main text, appendix </function>
    <format> same as dc:format, encoded in Internet Media Types, see IANA (2001)</format>
    <restriction> text that explains access to the file, as dc:rights </restriction>
<reference> a container for a reference made by the text
    <literal> text of reference </literal>
    <context> to context of the citation </context>
<type> a text type </type>
<identifier> an identifier for the text from a scheme that does not use AMF, as dc:identifier </identifier>

3.2.2: Text to person/organization verbs

<hasauthor> </hasauthor>
<haseditor> </haseditor>
<haspublisher> in the sense of dc:publisher </haspublisher>
<hassupervisor> </hassupervisor>
<hastranslator> </hastranslator>
<hasmaintainer> </hasmaintainer>

3.2.3: Text to text verbs

<iserratumof> </iserratumof>
<haserratum> </haserratum>
<isaddendumto> </isaddendumto>
<hasaddendum> </hasaddendum>
<isreviewof> </isreviewof>
<hasreview> </hasreview>
<iscommenton> </iscommenton>
<hascomment> </hascomment>
<istranslationof> </istranslationof>
<hastranslation> </hastranslation>
<isreplacedby> as dcq:isReplacedBy </isreplacedby>
<replaces>as dcq:replaces </replaces>
<ispartof> as dcq:isPartOf </ispartof>
<haspart> as dcq:hasPart </haspart>
<isreferencedby> for example cited by another, as dcq:isReferencedBy text </isreferencedby>
<references> e.g. cites another text, as dcq:references </references>
<isversionof> points to an earlier text that current text is developed from, as dcq:isVersionOf </isversionof>
<hasversion> points to a later text developed from the current text, as dcq:hasVersion </hasversion>
<isformatof> points to an original text with the same intellectual contents in a different format, as dcq:isFormatOf </isformatof>
<hasformat> points to a derived text with the same intellectual contents in a different format, as dcq:hasFormat </hasformat>

3.2.4: Text to collection verbs

<ispartof> a text belongs to a collection, as dcq:isPartOf </ispartof>

3.3: The collection noun

The collection noun element is used whenever statements about a set of several texts are being made. This can be a classification collection, (i.e. all the texts that have the subject classification code), a serial, the papers presented at a conference etc.:

<collection id="..."> ...adjectives and verbs... </collection> or <collection ref="..."/>

3.3.1: The adjectives of the collection noun

<title> same as a journal title, conference title etc. </title>
<abbreviatedtitle> abbreviation, e.g. PRL </abbreviatedtitle>
<description> a plain text description of the collection, as dc:description </description>
<homepage> URL for humans to read more about the collection </homepage>
<accesspoint> URL for machines to access the collection </accesspoint>
<type> a collection type </type>
<identifier> an identifier for the collection from a scheme that does not use AMF, as dc:identifier </identifier>

3.3.2: Collection to collection verbs

<isreplacedby> as dcq:isReplacedBy </isreplacedby>
<replaces> as dcq:replaces </replaces>
<ispartof> as dcq:isPartOf </ispartof>
<haspart> as dcq:hasPart </haspart>

3.3.3: Collection to text verbs

<haspart> as dcq:haspart </haspart>

3.3.4: Collection to p/o verbs

<haseditor> </haseditor>
<haspublisher> in the sense of dc:publisher </haspublisher>
<hasmaintainer> p/o who maintain metadata about the collection </hasmaintainer>

4: Element value constraints

The values of some of the elements are restricted. These content types are listed here.

4.1: The date type

The date adjective is of the form yyyy[–mm[–dd]], where [] encloses optional components. For details, see the date type definition of XML Schema.

4.2: The URL type

The value must be a valid Uniform Resource Locator.

4.3: The email type

The value must be a valid email address.

4.4: The collection type type

The collection noun covers a wide variety of things in AMF. It is useful to indicate the type of a collection through a controlled vocabulary.

book as OpenURL:book classification a classification scheme
proceedings conference proceedings series a series of texts
journal as OpenURL:journal archive an archive of documents

4.5: The text type type

The text noun covers a wide variety of things in AMF. It is useful to indicate the type of a text through a controlled vocabulary.

book as OpenURL:book article as OpenURL:article
conferencepaper as OpenURL:proceeding preprint also covers working papers and technical reports,
bookitem as OpenURL:bookitem as OpenURL:preprint

5: Attributes of elements

5.1: The id attribute

All AMF records (i.e. non-empty nouns) may have an id attribute. There are no value constraints for this attribute. If a value is set for a particular record, it is assumed that within the scope of a collection of AMF records, the record is uniquely identified by the value of this attribute.

5.2: The ref attribute

Any noun—be it empty or not—may carry a ref attribute. If it is present, its value is identical to the id attribute of another record. AMF ref attributes may be resolved to records that have the appropriate identifiers. The details of the resolution algorithm are outside the scope of AMF.

5.3: The from and until attributes

All verbs admit two additional attributes: from and until. The values of these attributes must be of the type date. These attributes indicate a time span for which the relationship holds. The dates are inclusive. Example:

<iseditorof from="1999-01-01" until="2000-01-01"> ... </iseditorof>

5.4: The xml:lang attribute

All adjective elements have an optional attribute called lang. It takes the same syntax as in the XML 1.0 specification. It uses values for xml:lang from

As a general rule, the lang attribute refers to the value of the element content. For example <title lang="fr">Robin des Bois<title> does not mean that the text is a French translation of the adventures of Robin Hood.

There are natural exceptions to this rule. For example, if the lang attribute appears at an adjective that is constrained to contain a URL, it means that text that is found at the URL is written in that language.

5.5: The event attribute

All date elements may have an optional attribute event that indicates what happened on the date. The admissible values are

created text was first written, as dcq:created or person was born, as vCard:BDAY
available date where a person was alive or a resource is available, as dcq:available
issued the formal publication date of a text, as dcq:issued
modified the date a resource was changed, as as dcq:modified

Example: <date event="created">2000-03</date>

5.6: The xsi:type attribute

The identifier, classification and keyword attributes may have an xsi:type attribute. In that case, controlled values for the value have been registered with AMF. The AMF Controlled vocabulary document lists all the controlled vocabularies.

6: Examples

None of the following examples is fictitious. However, the description of the items that is made through the examples may not be complete, to conserve space.

    <text id="bible"> <title key="Bible">The Holy Bible</title> </text>

    <text><title>The book of Genesis</title><ispartof>
        <text ref="bible"/></ispartof></text>

    <organization id="RePEc:edi:oecddfr">
    <name lang="en">Organization for Economic Development and Cooperation</name>
    <shortname lang="en">OECD</shortname>
    <name lang="fr">Organisation de Cooperation de
    Develloppement Economiques</name>
    <shortname lang="fr">OCDE</shortname>
    <homepage lang="en"> </homepage>
    <homepage lang="fr"></homepage>
    <haspart><organization ref="RePEc:edi:edoecfr">
        <ispublisherof><collection ref="RePEc:oed:oecdec">
                    <person ref="RePEc_per_1956-06-20_GIUSEPPE_NICOLETTI"/>
                The paper looks at patterns of regulation in service industries
                and explores their implications for service performance.
                <abstract lang="fr">
                Cette étude analyse les approches règlementaires dans les
                secteurs des services et explore leurs implications pour
                les performances sectorielles dans les pays de l'OCDE.

    <collection id="csfhrd">
    <title>Classification Scheme for Human Rights Documentation</title>
        <name>Ivana Caccia</name>
        <collection id="csfhrd:GEN II.10"><title>
        natural justice </title></collection>
        <collection id="csfhrd:GEN II.20"><title>
        universality / relativism </title></collection>
        <collection id="csfhrd:GEN II.30"><title>
        philosophy & human rights </title></collection>
        <collection id="csfhrd:GEN II.40"><title>
        political theories & human rights </title>
            <collection id="csfhrd:GEN II.41"><title>democracy</title></collection>
            <collection id="csfhrd:GEN II.42"><title>liberalism</title></collection>
            <collection id="csfhrd:GEN II.45"><title>marxism</title></collection>