Vous êtes ici

ISIDORE : Going from strenght to strength !

Version imprimableEnvoyer à un ami

In October 2009 TGE ADONIS launched implementation of the ISIDORE research platform, the object of which is to federate digital data from HSS research, thus providing a single access point for teachers, researchers, doctoral and junior students and also for third party web portals, using state-of-the-art data linking techniques.

What are we actually setting up ?
The ISIDORE research platform is different from conventional search engines in several respects :

  • Targeted harvesting of available resources freely accessible on the web ;
  • Indexing of non-structured data (full text, for instance) and of structured data (metadata, for instance) ;
  • Normalisation of metadata and enrichment of data by means of recognised authority files (PACTOLS thesaurus, RAMEAU vocabulary, HAL authors…) ;
  •  Graphical research interface making full use of the richness of structured data and of the authority files to provide an attractive and efficient user interface for the researcher
  • Highlighting indexed data sources ;
  • Availability of all enriched metadata through the search engine for general re-use ;
  • Possible integration of the ISIDORE search engine in another environment via standard web interfaces To do so, this research platform uses different ways of harvesting data, i.e. connectors :

- OAI-PMH warehouse harvesting ;
- Entry retrieval in trustee flows (RSS, ATOM) ;
- Targetted indexing of web pages on a site by means of both a “sitemap”-type XML flow and the presence of RDFa-encoded metadata.
These different connectors enable retrieval of a collection of metadata and form entry-points which complement the full text content of the page, which is also indexed when possible.
Enriched metadata, for their own part, are re-displayed according to the principles of web data governing the ways public data are made available such as data.gov and data.gov.uk. In this respect, the ISIDORE research platform constitutes a major contribution to the advance of HSS research in France. The actions carried out by the engine to support these functions are completely transparent, which is of particular importance since, in addition to user-friendlysearch interfaces, ISIDORE thereby provides means for the whole community to enrich community-held data constantly and consistently.
The search interface will also offer the ability to extend and complete a search across the combined bibliographical catalogues of other major data providers.
ISIDORE has been specifically designed to promote access to data generated by research organisations in higher education, laboratories, research teams : electronic publishing, documentary databases, digitised collections within research libraries, research diaries, and announcements of scientific events.

How far have we got ?
A significant development team has been created within the TGE ADONIS to construct this project, in collaboration with a service provider. The team includes a project management team (TGE ADONIS), an implementation team with a head of project (CCSD, the Centre for the direct scientific community, Lyon). The completed specification sheet was made public in July 2009 and available on-line on the TGE ADONIS web site. With the ongoing assistance of Atos Consulting and the selection of the service providers such as Antidot, Sword and Mondeca, the project has moved to the implementation phase. In November 2000, a first beta release of the platform will be open to the public. A general introduction is already available on-line.
In advance of the indexing of data within the platform, the TGE ADONIS has already initiated organisation of several meetings with major HSS data providers whose data will be accessible through ISIDORE. The aim of these meetings is to introduce the project and to facilitate working together on the harvesting side for improved indexing and presentation by the graphical interface. The following organisations have already been contacted : HALSHS, Revues.org, INIST, CAIRN, PERSEE, Gallica/BNF, etc. The following organisations will be approached in May/June : the Research Institute for Development (IRD), the Bibliographic Agency of Superior Education (ABES), National French Library. It goes without saying that a large number of data reservoirs (OAI-PMH, RDFa and RDF, RSS, Atom) other than those generated by these producers will be harvested by ISIDORE (BU, research libraries and documentation centres, research centres, Digital Resources Centres, etc.) The open and open-ended list is accessible on-line. Please do not hesitate to contact us for more information.
Beyond the technological innovation and the perspectives of the data web, ISIDORE is above all a remarkable federative effort on the part of a very large number of data providers, documents, and sites. It is not simply a question of complying with the same standards, but also uniting in the interest of all parties to support the unimpeded development of usages and services for the benefit of the whole French or French-speaking community in the human and social sciences. In this sense, ISIDORE is a major constituent of the distributed access architecture which the TGE ADONIS has been setting up for the last three years. As regards the implementation, the service provider is making good progress on the three components of the platform : collection, processing and access.

Collection
The different connectors described previously are currently being finalised following development of appropriate criteria for documenting each type of data source (site URL, warehouse URL, date format…) using an ad hoc graphical interface.

Processing
ITM, the reference management software supplied by Modeca, is in place on the ISODORE computer platform at CCSD. The various authority files contemplated as of the origin of the project are already integrated or in the process of being integrated (Rameau, Pactols, scientific domain of the HAL open archive, the HAL authors…) and we are currently working on TermSciences (INIST). Tools for normalisation of dates and personal names are being finalised and the data model has now been completed.

Access
Web interfaces are being defined by the CCSD. Three interfaces will be available : via the TGE ADONIS site, via a generic site in progress, via widgets and modules for the partners.
To support access and display of or data according to the data web principles, if the data provider of resources indexed within ISIDORE does not currently supply a normalised, reliable, and platform independent identifier (ARK, DOI) , one will be allocated according to the HANDLE principle. Whether to retrieve such identifiers and to use or expose them rests with the data provider.

What remains to be done ?

  •   New meeting with ABES (authority file for authors, further research on SUDOC from ISIDORE)
  • Publishing the guide on OAI-PMH for laboratories and « How to be harvested and indexed by the ISIDORE search engine ? ».
  •   Setting up a prototype (alpha release) in June 2010 intended for the first producers (see above).
  • Realising the web accesses and the widgets (July/September 2010). Setting up a work group on the use of authority files in ISIDORE (July 2010).
  • Development of the sitemap/RDFa connector (September 2010).

What future prospects are envisaged ?

This ISIDORE access platform, by mobilising the data, metadata and authority files of numerous institutions and various disciplines in HSS, higher education and research, aims to be a shared and open data warehouse from which a very large number of services and disciplinary portals may develop freely on the web. It will make possible an exponential increase in the richness and quality of French scientific materials and productions. Consequently, this first step, the public launch in November 2010, will be a key starting-point for future extensions of the scope of the platform to include many other appropriate data sources, not excluding data in other languages with the possibility of multilingual indexing, within the framework first of all of the ESFRI roadmap and of its DARIAH, CLARIN or CESSDA projects and also for accommodating new shareable services which may or not rest on data provided by ISIDORE within the ADONIS infrastructure deployed at the CCSD.

Laurent Capelli, Jean-Luc Minel, Stéphane Pouyllau, Gauthier Poupeau, Jean-Baptiste Génicot, Yannick Maignien