Archive for January, 2007

How CIRAD stands vis-à-vis the Open Archives projects


At the local level

CIRAD’s database, AGRITROP

AGRITROP is CIRAD’s bibliographic and full text database.


AGRITROP is composed of references associated with full text (when it exists) of all CIRAD’s publications (books, book chapters, conference proceedings, conference communications, journal articles, training courses, pictures) and of the books available in CIRAD’s libraries.


AGRITROP uses the AGRIS/CARIS categorization scheme and the French AGROVOC thesaurus to index its data.


AGRITROP is based on the library information system LORIS/DORIS which is a commercial software developed by a private French company EVER. LORIS/DORIS is not OAI compliant.

UNIMARC is the format of AGRITROP.


Outputs from AGRITROP

CIRAD hasn’t supplied AGRIS with AGRITROP data for several years.

Because AGRITROP isn’t OAI compliant, CIRAD has to extract publications lists from AGRITROP and then display them on the Internet ( so that references and full text can be indexed and retrieved by search engines like Google and Google Search.


Data from AGRITROP could be exported in a XML format but the output and process are complex and can’t easily be used to export data according to DC or AGRIS AP.



At the national level

CIRAD’s participation in the national Open Archive

CIRAD has been participating since 2006 in the national Open Archive HAL (Hyper Article on Line) which is hosted and managed by CNRS (the National Scientific Research Center).

HAL was initially designed to host scientific publications which are accepted by peer-review journals.

HAL covers a wide range of scientific areas from mathematics and physics (HAL is linked to Arxiv) to medicine (HAL will soon be linked to PubMed), psychology and agronomy.

HAL is harvested by Google and Google Scholar.


Our main concern with HAL is that it is not intended to host unconventional literature (technical documents, activity and expert reports, unpublished communications). Its classification scheme isn’t a conventional one and its keywords aren’t controlled.



CIRAD’s database isn’t OAI compliant and is not directly available through OAI harvesters and search engines like Google and Google Scholar.

Although CIRAD participates in the national Open Archive HAL, it can’t use HAL to promote CIRAD unconventional publications which form the greater part of CIRAD’s literature.


Furthermore, the agronomic field is a minor subject in HAL which isn’t based on a well established classification scheme (like the AGRIS/CARIS) nor as a controlled vocabulary (like AGROVOC).



CIRAD’s requirements in order to participate at the international level

The best option for CIRAD, in terms of access to its publications, would be to set up an institutional repository based on OAI standard and controlled vocabularies (ISI subjects for scientific classification, AGRIS/CARIS for applied research fields and AGROVOC for keywords).

This repository would accept any type of publication (from peer review articles to technical documents) in order to give access on the Internet to unconventional literature produced by CIRAD’s researchers.


What CIRAD thinks of AGRIS AP

It is essential for our repository to use a standard and as complete as possible a metadata set. Moreover, it is essential that CIRAD’s publications can be harvested, searched, retrieved and read on the Internet according to standard data. However, AGRIS AP is a little too heavy and complicated for exporting data easily in a XML format.

What we need is a simplified metadata set from AGRIS AP which would be processed by service providers, like the one FAO proposes to set up for the AGRIS network and the international agronomic research community.


MC Deboin 19/01/2007




Leave a Comment

Benefits of AGRIS AP over simple DC in OAI environment

In the OAI scenario, to establish good services we have to keep in mind the “garbage in, garbage out” principle. Quality services are those that provide the users with good search and browse functionalities, where the user is served as precise results to her query as possible. Given that we try our best in our own environments (individual organizations) to create best quality metadata, it becomes important to assure a level of quality in a shared environment.

I am posting here a document that argues that to have high quality value-added services, we need more than just simple Dublin Core and that a possible solution could be the AGRIS Application Profile.

Read more: Benfits of AGRIS AP over Simple DC 

Gauri Salokhe

Comments (2)

Draft Agenda for AGRIS Content Management Taskforce

The draft agenda for the meeting of the AGRIS content management taskforce has been published at All task force members, may comment the draft agenda

Leave a Comment