The International System for Agricultural Science and Technology (FAO AGRIS) has, for more than five decades, played a central role in facilitating access to agricultural scientific and technical information. Since its establishment, FAO AGRIS has supported contributing partners in creating and submitting bibliographic metadata to enable the discovery, exchange, and reuse of agricultural research outputs.

Over time, the FAO AGRIS metadata framework has evolved significantly. Early approaches were based on centrally coordinated guidance for metadata creation and submission. From the early 2000s onwards, a major conceptual and technical shift took place with the development of the AGRIS Application Profile (AGRIS AP). This profile was designed to improve interoperability and metadata quality by combining elements from widely used standards, including Dublin Core, AgMES, and the Australian Government Locator Service (AGLS). The AGRIS AP provided a structured approach for describing Document-Like Information Objects (DLIOs) and enabled consistent metadata exchange across distributed systems.

A key milestone in this evolution was the Consultation on Agricultural Information Management (COAIM, 2000), which recognised FAO AGRIS as a potential catalyst for a new model of agricultural information management. The consultation emphasised the dual role of FAO AGRIS as both a capacity-building initiative and an information system. Subsequent discussions, including COAIM 2002, reinforced the need for a common metadata standard to improve interoperability, support harvesting, and enable cross-searching across heterogeneous repositories.

With the introduction of interoperability protocols such as OAI-PMH (Open Archives Initiative Protocol for Metadata Harvesting), FAO AGRIS transitioned towards a more decentralised network model. This shift allowed a broader range of institutions, not only established FAO AGRIS Resource Centres but any agricultural research organisation, to contribute content to the system. These institutions, acting as data providers in a distributed environment, may expose metadata in multiple formats, including AGRIS AP, Dublin Core, and others, thereby increasing visibility and dissemination opportunities.

However, this increased flexibility has also introduced variability in metadata quality and consistency. While openness has strengthened participation and increased content volume, it has also created challenges in harmonisation across key metadata fields. Among these, the Type of Document element has been identified as particularly critical, as inconsistent usage has led to classification variability and the inclusion of records that are not always relevant to FAO AGRIS scope.

Accurate document type classification is critical for search filtering, discovery, analytics, content management, and quality assurance. Within FAO AGRIS, inconsistencies in the use of document type values across data providers have affected metadata quality and reduced the precision of search and filtering functions. To address these challenges, the Type of Document Vocabulary (ToDVoc 1.0) has been developed as a controlled vocabulary to standardise document type assignment, strengthen metadata quality, and support more effective access to agricultural information.

The objectives of ToDVoc 1.0 are to:

● Provide a controlled and standardised vocabulary for document type classification within FAO AGRIS.

● Improve metadata consistency and quality control across distributed data providers. 

● Support the identification and filtering of relevant content, ensuring alignment with FAO AGRIS scope.

● Enhance interoperability across multiple metadata standards and formats (e.g. AGRIS AP, Dublin Core, MODS, EndNote, DOAJ). 

● Improve searchability and discovery of FAO AGRIS records through harmonised classification.

● Support multilingual consistency and future translation of document type terms.

● Strengthen metadata governance in a decentralised harvesting environment.

General Principles

ToDVoc 1.0 is the authoritative reference for assigning document types in FAO AGRIS. It must be used to ensure consistent classification across all records, regardless of data provider or metadata format.The vocabulary is intended to reduce ambiguity and variation in document type assignment and to ensure alignment with FAO AGRIS quality and scope requirements.

Assignment Rules

● Each record should be assigned a single primary document type that reflects its main intellectual content.

● The assignment must be based on the content of the resource, not its technical format (e.g. PDF, webpage, presentation slides).

● When multiple interpretations are possible, the most specific and appropriate term from the vocabulary should be selected.

● Free-text or locally defined document types must be avoided when a controlled term exists.

Scope and Quality Control

In a decentralised harvesting environment based on OAI-PMH and multiple metadata formats, variability in metadata quality is expected. ToDVoc 1.0 supports improved quality control by:

● Enabling identification of non-relevant or out-of-scope records.

● Supporting consistent classification and reclassification of incoming metadata.

● Improving the overall quality and coherence of the FAO AGRIS database.

Maintenance and Updates

ToDVoc 1.0 is maintained by the FAO AGRIS team and will be updated periodically.

Updates may reflect:

● feedback from data providers and users,

● changes in scholarly publishing practices,

● emerging document types,

● and developments in metadata standards.

ToDVoc will be translated into FAO's official languages to support multilingual discovery while preserving a single controlled concept for each document type.

The vocabulary will be maintained with a focus on stability, interoperability, and backward compatibility.3.5 Compliance with ToDVoc 1.0 (Data Provider Requirements) Data providers are strongly advised to supply records with document types mapped and aligned with the controlled list defined in ToDVoc 1.0. The use of controlled document type values is a mandatory requirement for inclusion in the FAO AGRIS system.

Records that cannot be mapped to an accepted document type aligned with ToDVoc 1.0 will be excluded from FAO AGRIS. However, exceptions may be considered when data providers provide a valid justification for the inclusion of a new document type in FAO AGRIS. When the same concept can be expressed using different labels, it may require mapping, normalisation, or further review during ingestion and quality control processes.

This includes variations such as translations, misspellings, abbreviations, and other alternative forms of the same term. In particular, records corresponding to ephemeral, non-scientific, or non-technical materials are not accepted within FAO AGRIS and will be excluded. This includes, but is not limited to:

● news bulletins and press releases

● blog posts and blog entries

● editorials and opinion pieces

● newsletters of non-scientific nature

● announcements and event advertisements

● flyers, leaflets, brochures, and promotional material

● institutional web updates without research or technical content

● other short-lived or non-permanent informational content

To ensure successful ingestion and long-term availability of records, it is strongly recommended that all submissions include a valid document type from ToDVoc 1.0, or a local document type that can be clearly mapped to a ToDVoc term. Records submitted without a document type, or with uncontrolled values, may be rejected or excluded from the FAO AGRIS database.

The introduction of ToDVoc 1.0 represents a key step in strengthening metadata consistency, quality, and interoperability within the FAO AGRIS system. In a distributed and multi-standard environment, a controlled vocabulary for document types is essential to ensure reliable classification, improved discoverability, and effective access to agricultural knowledge worldwide. ToDVoc also provides the document type labels used in ODS, ensuring alignment between systems and supporting more consistent metadata practices across data providers.

Table 1. ToDVoc 1.0