FAO has a normative role to play in the creation and promotion of standards in agricultural information management. It understands itself as an organization responsible for knowledge brokering. Knowledge and information can be brokered on vertical or horizontal lines only when common exchange standards exist. Before the explosion of the Web this problem had been mainly handled by the international Library community and its Z39.50 protocol [12] for information exchange.
In 1995, the Online Computer Library Center [13] organized the first workshop of the Dublin Core community with the aim of reaching a consensus on a core set of metadata elements to describe networked resources. This discussion was taken up in a meeting on agricultural standards, organised by Oneworld Europe [14] in collaboration with FAO in Brussels, during autumn of 2000 [15]. The meeting raised awareness amongst information providers of the new opportunities for sharing information through use of metadata standards and platform-independent formats such as XML. The participants agreed that a standard methodology was necessary and showed particular interest in the Dublin Core model.
Since the Brussels meeting in 2000, FAO has undertaken a number of initiatives to facilitate the standard setting process. More recently, in June 2003, the United Kingdom Department for International Development [16] organized the “Fertile Ground” meeting, where policy makers and information service managers discussed the implications of the dispersed developments in exchange standards, tools and methodologies, thus introducing the notion of - and the need for - “coherence”. This coherence can be achieved by: facilitating collaboration, partnership and networking among partners by promoting information exchange and knowledge sharing; and by harmonizing the decentralized efforts currently taking place in the development of methodologies, standards and applications for management of agricultural information systems; consequently, providing a ‘one-stop’ access to system designers and implementers.
The Agricultural Metadata Standards Initiative (AgMES), launched in November 2000, is an attempt to promote the use of metadata through use of standardized agricultural metadata terms for the purpose of facilitating resource discovery and interoperability between richly described agricultural resources. It defines elements, qualifiers, encoding schemes and controlled lists that are necessary for the description of agricultural information resources. The use of such a standard makes it easier to integrate data from different sources allowing for creation of value-added services such as simple aggregated subject-based views, automatic news feed services etc.
Much of the activity has taken place in collaboration with internal and external partners. The AgMES Web site has been continuously updated with useful information for the users. It is now furnished with official documentation, metadata schemas, metadata creation tool, glossary and FAQs. A mailing list [17] provides a forum for discussions and exchange of ideas. The AgMES initiative has received encouraging feedback internationally with a number of citations on well-known sites. Additionally, conference presentations and journal articles have been written on the topic [18, 19, 20]. Several applications, both within and outside FAO, have fully or partially implemented metadata schemas using AgMES elements (e.g. Open Knowledge Network [21], Consultative Group on International Agricultural Research [9], FAO Fisheries Global Information System [22], International Portal on Food Safety, Animal and Plant Health [23], and Global Forest Information Service [24]).
One of the first AGMES applications was the AGRIS application profile [25]. AGRIS is the international information system for the agricultural sciences and technology. It was created by FAO in 1974, to facilitate information exchange and to bring together world literature dealing with all aspects of agriculture [26]. The AGRIS Application Profile is a format that allows sharing of information across dispersed bibliographic systems. Most information systems, including AGRIS, are faced with the chronic problems of exchanging and aggregating information, starting from the differences in applications to those concerning varying database structures and cataloguing rules. The AGRIS Application Profile is a major step towards exchanging high-quality and medium-complexity metadata in an application independent format and provides possibilities to offer value-added services, irrespective of how the information was stored locally.
The following figure shows how the use of an Application Profile as a common exchange layer helps to resolve the problem of heterogeneity of data models among information systems with similar types of resources (i.e. Database 1, Database 2,... ), and can be used as a basis for the development of further value-added services (i.e. Thematic Portals, News feed Services,... ).
Figure 2. Interoperability, using common exchange standards, between distributed datasets allow for creation of value-added services
The result of introducing this standard can be measured quantitatively. This year, so far in the first nine periodical updates, 44105 records have been added; a significant increase from last year’s total of 38532 records that were uploaded. One third of these records were submitted using the AGRIS Application Profile XML format and WebAGRIS (as it is now available with the automatic XML-based export functionality). The most remarkable achievement of the new format is resuming of records submission by some of the countries that had stopped sending in their inputs to the AGRIS database, such as Finland and Norway.
The multilingual AGROVOC thesaurus [27] is available in the five FAO official languages which are English, French, Spanish, Chinese and Arabic. It is also available in Czech and Portuguese and Thai. Other languages such as Italian, Korean, Japanese, Hungarian and Slovak, are currently under translation and revision. The main role of AGROVOC is to standardize the indexing process in order to make searching simpler and more efficient, and to provide the user with the most relevant resources. Currently, it is downloaded regularly, an increase possibly attributed to its ease of applicability in multilingual web applications.
AGROVOC is the foundation that underpins the development of the Agricultural Ontology Service (AOS) project. By making use of knowledge contained in vocabulary systems and thesauri such as AGROVOC, AOS is committed to developing specialized domain-specific terminologies and concepts that will better support information management for the web environment.
Since Tim Berners-Lee has launched the debate about a future “Semantic Web” [28], there is a growing interest in vocabularies and especially in structured controlled vocabularies among the community of Web Information Specialists. FAO has taken up the challenge by re-launching the AGROVOC thesaurus as a starting point for a basis for using ontologies in Agricultural knowledge applications. The first AOS Workshop [29], which was held in 2001 in Rome, was aimed at transforming the AGROVOC Thesaurus into an Ontology server. Since then the approach has become much more phased and is directed at developing a collection of semantic tools which will serve as building blocks of the AOS project. Since 2001, five Agricultural Ontology workshops have taken place, allowing experts to exchange knowledge about the semantic tools and trends and to raise awareness of its importance for the Agricultural community.
Using AGROVOC, Domain Ontologies [30] have been developed and are currently being used in some of FAO’s applications [31]. Further funds and collaborations are needed to extend the work in this area. A number of conference presentations and journal articles have been written on the topic [32, 33, 34, 35, 36, 37, 38, 39]. A recent publication also outlines the steps that will be taken towards removing inconsistencies from the thesaurus and enforcing on it a more ontological structure.
Although much has already been done in the development of information management standards, the outreach within the sector is still limited. A more solid framework needs to be adopted to raise awareness and promote ‘reuse’ of such standards in the development of new information services. The aim is to establish an Information Management Clearinghouse (IM Clearinghouse) to bring together information on methodologies, standards and tools. Its specific goals would be to:
establish a network of partners who adhere and agree to the primary goal of the initiative “to provide unified and free access to information management approaches and tools”;
bring together information about currently available standards (such as thesauri, classification schemes, metadata sets, ontologies, controlled vocabularies) used in management of Agricultural Information;
encourage the re-use of these standards to facilitate interoperability between information systems;
increase awareness of these freely available resources and promote sharing of good practice examples; and
provide channels for communication between different actors in the community.
Figure 3: Adoption of standards from the Clearinghouse will facilitate interoperability between systems and allow for creation of value-added systems.
In Figure 3, the lower layer depicts the different collections of data (ranging from numerical datasets to collections of full text documents). These collections are compiled by agricultural information providers, but not necessarily by the same services that provide the final views to the community. The top layer is made of different Agricultural Information services, such as multi-host database searches, e-mail alert services about new outputs, portals, or paper-based journals. The proposed intervention of the Information Management Clearinghouse, to facilitate interoperability, would take place in the middle layer, where developers of information services will be able to obtain information on methodologies that have been applied, the standards that have been used and the way they have been applied. This framework of sharing common standards and tools would not only influence both the data-providers and service-providers.
The increasing uptake of common standards will provide the possibility not only to enhance interoperability but also to decrease the cost of the resources necessary to create information systems from scratch and the reuse metadata. This is clearly exemplified in the case of bibliographic agricultural information systems such as AGRIS. Although existing systems have started to share information using standardized data-models and XML technologies, this is not the most efficient solution for systems that are being newly created. An acceptable level of homogeneity within the domain can be achieved by sharing not only ‘transformed and compliant’ metadata but also by sharing the backbone technologies and tools used to create those information systems.
The scope and eventual sustainability of the Information Management Clearinghouse will be determined by the level of cooperation received from the participating network of partners.
Article 1.1 of FAO’s constitution states the “the organization shall collect, analyze, interpret and disseminate information relating to nutrition, food and agriculture". Moreover, as mentioned in section one, of the key strategies of FAO is to improve decision-making through the provision of information and assessments and fostering of knowledge management for food and agriculture. In support of these critical functions, FAO established the World Agricultural Information Centre (WAICENT) [4] as a corporate framework for agricultural information management and dissemination.
One of the fundamental principles of WAICENT is that is based on a participatory and decentralized approach. Each technical unit is responsible for collecting, analyzing, interpreting and disseminating the information for which it is responsible. The amount of information and knowledge acquired, managed and published under this approach is enormous as manifested by the half-million HTML pages available on the corporate web-site. However, at the same time, FAO strives to provide an “integrated information resource base, with current, relevant and reliable statistics, information and knowledge made accessible to all FAO clients” [2]. Without the guidelines and standards offered by the WAICENT framework, it would be virtually impossible to obtain integrated access to FAO’s information under a decentralized approach and it would be very difficult to cover the many interdisciplinary and cross-cutting themes that FAO is mandated to address.
The vast amount of information generated by a large organization such as FAO, together with the apparent dichotomy between decentralized production on one hand, and a desire for integrated and interdisciplinary information access on the other, makes the WAICENT Framework a rather complex institutional arrangement. For WAICENT to work efficiently and effectively, all FAO technical departments are involved in the policy-making, approval and implementation process through two corporate WAICENT bodies: the Committee and the Advisory Group. Adoption of standards would not be possible in a decentralised environment like FAO if broad participation of information originators were not ensured. Within the Organization, the WAICENT Framework acts as a sounding board and policy-making mechanism for the discussion, adoption and implementation of standards.
The WAICENT Committee works to enhance the sense of ownership of WAICENT by all organizational units concerned and bring about a shared understanding of how WAICENT can better serve the needs of FAO target audiences. It also recommends policy decisions which are then passed to the FAO Director General for final approval.
The WAICENT Advisory Group (WAG) supports the WAICENT Committee by proposing items for the latter's agenda, gathering and preparing necessary background information for agenda items, tracking and reporting on the implementation of the Director-General's policy decisions. The WAG also resolves many technical issues and ensures that standards and policies for development of interoperable systems are based on common standards, tools and methodologies.
