ERCIM News No.28 - January 1997

Calliope - An Experiment in Digital Libraries

by Catherine Alauzun and François Rechenmann

Calliope is a joint project between INRIA Rhône-Alpes, Imag and Rank Xerox Research Center, all of them located in Grenoble, France. It aims at providing the researchers of these three laboratories with document image storing and retrieval facilities, together with a 'scanning on demand' service, thus allowing them to share their document resources. The project has been motivated by the location of the newly created Inria Rhône-Alpes research unit some kilometres away from the 'médiathèque Imag' which has accumulated and organised document resources in computer science and applied mathematics over the last forty years. The founding principle of Calliope was thus to give researchers at INRIA Rhône-Alpes a private access to part of these resources using network facilities.

Calliope is based on the XDOD system (Xerox Document On Demand) developed by Xerox, for the storing and retrieval of the document images, and its Web interface called DocuWeb. Calliope makes use of a list of serial titles together with the contents of each of their issues. This list is weekly updated from the SwetScan server located in Amsterdam. The user can browse through these contents and select papers of interest.

Once a paper has been selected, a copy can be ordered by e-mail. If the issue is available on one of the three sites, it can be scanned and then stored on the Calliope server. The researcher can then read the paper on the screen of his/her workstation with a low level quality, or print it on the nearest printer. Obviously, if the same paper is requested by another researcher, it will be available at once, without any scanning procedure. If the issue is not available in any connected site, a classical paper copy can be requested.

The expected advantages of Calliope over a traditional service of article photocopying can be evaluated from several points of view: the user/researcher, the manager of the documentation centre, and the publishers.

The researcher will first benefit from an easy and fast access to the information. But (s)he will also appreciate the browsing and request facilities, such as keyword-directed search on the titles of the articles. Moreover, a subscription service is provided: (s)he can ask to receive directly by e-mail the contents of the newly arrived issues of the journal titles (s)he has selected.

The manager of the documentation centre will appreciate the savings which result from the fact that the scanning process, more or less equivalent to the photocopy process in terms of requested manpower, can be written off. In a traditional photocopying service, users may indeed ask for copies of papers after consulting the contents, generally available on a paper support. As with Calliope, the criteria for selection are the title of the paper and the names of the authors. One of the main advantages offered by Calliope is that, if the article does not match the user's expectations, the scanning efforts will perhaps benefit to another researcher later. In the same line of reasoning, any paper which has been scanned might be immediately read or printed by another user.

Calliope is designed to keep tracks of every operation and maintain detailed statistics, such as the number of pages scanned, accessed and printed by the users. These statistics will obviously help to preserve the publishers' interests.

Calliope is presently under experimentation. Simultaneously, legal problems are asserted and solutions for preserving the rights of the publishers are being investigated. Only when these problems are solved, the Calliope project will evolve as a fully operational service.

Technical extensions are already anticipated. In the present architecture, all document images are stored on only one server. A first extension will thus consist in managing several servers, for instance one for every scanning site. Turning images into texts through OCR will offer new capabilities, such as full text searching, automatic generation of abstracts or translation assistance. The Navibel prototype designed by Rank Xerox Research Center exemplifies some of these extensions. It allows the user to add annotations to a scanned document, create links to other documents and manage private libraries.

Info on the web:

Rank Xerox Research Center:

Médiathèque Imag:

Please contact:
Catherine Alauzun - INRIA
Tel. +33 4 76 61 53 10
E-mail :

return to the contents page