Metadata harvesting is at the heart of the International System for Agricultural Science and Technology (AGRIS). Managed by the Food and Agriculture Organization of the United Nations (FAO), AGRIS aggregates thousands of bibliographic records each month from more than 1,600 data providers worldwide. AGRIS harvests metadata through the Open Archive Initiative-Protocol for Metadata Harvesting (OAI-PMH). 

OAI-PMH defines the technical specifications necessary for the exchange of metadata between a data provider and a data harvester. Through this protocol, metadata is exposed on the World Wide Web, organized in sets. The classification of the sets is important to help harvesters select relevant metadata of research results, data type, or scientific publications to display on their platforms. 

AGRIS is a mature system with long-term continuity, with 50 years of experience in collecting metadata from data providers all over the world. Based on this extensive experience gained from the collection of millions of records and the technical knowledge of FAO experts, the following good practices have been identified.

Organize Sets by Subject or Theme, or Both

Structure your OAI-PMH sets around thematic areas relevant to agriculture (e.g., crop science, soil management, livestock). Avoid organizing exclusively by document type, for example, or just organizational units in your institution with varying document types.

Subject sets improve the ability of AGRIS and any other subject service provider to identify, filter, and index content effectively, improving relevance and discoverability.

Use English Language to Label Sets

Ensure that set names are in English or languages that are widely used, to facilitate rapid identification of the sets to be collected, enabling global interoperability. Set labels in other languages that are not widely popular may be misinterpreted, or harvesters may overlook sets for data capture, reducing the likelihood of accurate collection.

Facilitate Access, Avoid Access Restrictions

Your OAI-PMH endpoint must be publicly accessible, without authentication (e.g., usernames, passwords...) or IP-based restrictions. AGRIS cannot collect data from non-publicly accessible repositories.

Apply UTF-8 Encoding Consistently

UTF-8 ensures that special characters and multilingual metadata are displayed correctly. Incorrect encoding may result in unreadable records during data ingestion.

Ensure Metadata Quality by Using Minimal Metadata Elements

Producing and exposing metadata on the web is an expensive exercise. That is why it is very important that cost is closely related to efficiency. By publishing metadata in any form and at any cost, your publications will not be more visible. It is advisable to consider metadata that will provide a complete record suitable for easy identification on the web. For this purpose, the minimum metadata requirements for AGRIS are recommended.

Use Persistent Identifiers (e.g., DOIs, Handles...)

Stable identifiers provide long-term access to full-text content. Broken or unstable links diminish the utility and reliability of your repository entries.

Validate EndPoints Regularly

Use external validators to periodically check your OAI EndPoint. This helps to identify problems in advance and ensures that metadata is always ready for harvesting.

Plan and configure the list of sets in advance

It is recommended to think in advance about what set list will be needed to facilitate the identification and harvesting of your metadata and, even if no records are available yet, to configure the sets appropriately from the start. Typically, harvesters access the OAI EndPoint at the start, when they have identified an information resource with records relevant to their purpose. Unfortunately, an unconfigured or undefined set specification may prevent service providers from looking at the relevant content available in their repository.

Regularly updating your records and maintaining a consistent metadata structure maintains content quality and significantly amplifies your research's global visibility and impact. Properly structured metadata encourages collaboration and positions your institution as a valuable contributor to an open, accessible, and dynamic agricultural knowledge ecosystem.

Get Support and Stay Connected 

Institutions seeking further guidance can enroll in the self-paced AGRIS Data Provider Online Course (available in English and Spanish), consult the detailed AGRIS User Guide and Data Provider Guides, or contact AGRIS directly for personalized assistance at [email protected]. Additionally, we encourage institutions to stay informed through our website, follow updates on our X account @FAOAIMS, and subscribe to the AGRIS Newsletter.