Fabrizio Celli (Software Engineer) is the technical leader for the AGRIS team and since February 2010 he has been working on the improvement of AGRIS developing the version 2.0 of the AGRIS search engine, and dealing with Linked Open Data technologies. He has been leading the developments towards the AGRIS 2.0 and has been instrumental in development of AGRIS and the systems behind it. We took this opportunity to ask him a few questions on AGRIS and its future developments.
1. Could you briefly tell us something about the history of AGRIS; why and how it came to be from its start in the 1970s until the release of AGRIS 2.0 last December 2013?
AGRIS is an initiative that was set up by FAO of the United Nations in 1974 to make information on agriculture research globally available. The AGRIS portal is one of the most important FAO web sites with an average 350,000 visits per month. The historical overall objective of AGRIS is to improve access and exchange of information serving “the information needs of developed and developing countries on a partnership basis”. Over 150 participating data providers located in more than 100 countries are currently aggregating scientific and research publications, scholarly papers and grey literature not officially published in commercial channels.
In order to cope with its goal of improving access to information in the agricultural domain, AGRIS adopted many technologies during its life: for instance, the AGRIS database changed from a relation database, to a filesystem XML database, to a triplestore, which is the technology currently used by AGRIS. These transformations highlight the change of perspective related to the AGRIS repository: originally the value of the system was only related to its precise and selected content (i.e. the bibliographic database in agriculture); now the focus is both on the AGRIS database and on external information that can be retrieved starting from the AGRIS core database.
Implementing the database as a triplestore allows the application to use the Linked Open Data power, so AGRIS 2.0 uses the technology to retrieve as much information as possible about a specific topic. The user searches for some keywords and retrieves some documents from the AGRIS database: AGRIS queries external sources of information to display other valuable information about the same topic (publications, maps, statistics, and pictures).
2a. AGRIS consists of a network of more than 150 centers from 100 countries. What is the role of these centers?
Traditionally, traditional AGRIS centres (national libraries) sent periodically their data to the AGRIS Secretariat to have them published in the AGRIS database. In the last years, not only traditional AGRIS centres but also journal editors create their metadata to be published in the AGRIS database and, with the growth of open access institutional repositories, AGRIS has dramatically improved its methods for harvesting and indexing metadata from content providers.
2b. How can other data providers, not yet involved, collaborate in AGRIS?
The OAI-PMH protocols currently allow the AGRIS service to facilitate the interoperability with service providers, by means of a more direct channel of communication between the two parties. This means that now the name “AGRIS centers” stands for all the data providers that have contributed or are contributing with bibliographic data to AGRIS, from national libraries, institutional repositories, single or corporate journal publishers to service providers.
If other data providers want to start a collaboration with AGRIS, they can read the documentation available in the AGRIS portal. AGRIS tries to accept metadata in many formats and suggests many tools to create metadata.
3. Could you tell us something about the main challenges you have encountered?
Challenges can be discussed focusing on different categories: the definition of the vision; the implementation of the vision (which includes the choice of the technology); the communication with users having different backgrounds and cultures; the collaboration with data providers, partners and colleagues; the responsibility of developing a big and important system, used by many users all over the world.
Anyway, as architect of the system, I can say that the main challenge is the understanding of the requirements of an international community, especially regarding how data can be exposed, interconnected, and used. Coping with many users with different backgrounds requires mental openness to really understand and quickly interpret their needs, their feelings, and the way they work and approach the usage of information.
4. What are the expected developments regarding AGRIS for the near future?
Our vision is to make AGRIS a social portal in agriculture. This means a place where users can register, can customize the mashup page pre-selecting external sources of information they are interested in, can discuss topics and research areas, and so over. This is a big process which requires many sub-phases, in the spirit of AGILE development: these sub-phases are not static, but they can evolve depending with new needs and requisites discovered during the development of the portal.
Questionnaires and interviews will be submitted to users with different backgrounds (librarians, researchers, cataloguers, software developers, students, etc.) to get their feedbacks and suggestions about the current implementation of the system and future desires.
Next steps will be the revision of the AGRIS backend, to have a single database (currently filesystem XML database and triplestore) and a data flow architecture much smoother, users registration and customization of the mashup page introduction of new services, also considering users feedbacks.