New IR Software: Archimede

The following comes directly from the Archimede website:

Laval University Library recently launched the third component of its institutional repository. Called «Archimede» (http://archimede.bibl.ulaval.ca), this component covers e-prints, pre-prints, post-prints and other research publications from faculty members and research communities.

Following a thorough analysis of available software solutions, including E-prints and D-Space, the library decided to develop its own customized application. Inspired by the D-Space model, Archimede is arranged around research communities and fully developed in open source. The system is OAI compliant, using a Dublin Core metadata set. An open source distribution of Archimede will be available soon.

Following are some highlights of the special features and characteristics of the system:

• Archimede has been developed in a multilingual perspective, with internationalization as a focus. Using the open source standard (i18n), the text (or content) of the interface is independent and not embedded in the code. It is then relatively easy to develop an interface in a specific language without having to work on the code itself. English, French and Spanish interfaces are already offered in Archimede. That feature allows also the user to switch easily from language to language anywhere and anytime during his search and retrieval process.


• Archimede is flexible and not dependent on a specific platform. The system can be installed on Linux as well as on Windows. For a library wishing to implement the solution, the system can be easily adapted to the technical infrastructure already in place, thus increasing the efficiency of the implementation process.


• Archimede allows searching on metadata as well as on the full text, thus enhancing the power of the search engine. An application is being developed that will automatically generate and translate from the text and the abstract a proposed set of controlled vocabulary subject headings. This will be done through the « Répertoire des vedettes-matières de la Bibliothèque de l'Université Laval » and its links to LC Subject headings, Canadian Subject headings, Mesh hand AAT.


• The search engine is based on open source Lucene, using LIUS (Lucene Index Update and Search), a customized framework developed at Laval by the library staff. LIUS allows indexing of different types of documents formats : XML, HTML, PDF, RTF, MS Word, MS Excel, JavaBeans; it also permits mixed indexing, integrating for example in the same occurrence metadata in XML and full text in PDF, HTML, etc.