User's Behaviour inside a Digital Library

Introduction

The CASPUR Consortium was established on June 5th, 1992; its name comes from the acronym: Inter-University Consortium for the Application of Super-Computing for Universities and Research. The Consortium headquarter is in Rome, Italy.

CASPUR is a no-profit Organization; it is financed by MIUR (the Ministry for Education, Universities and Research) and by associated Universities (mainly located in the Centre-South of Italy).

CASPUR main purposes are:

• •

  To manage a center capable of guarantee a high quality and high-powered processing service;

• •

  To promote the use of the most advanced information processing systems;

  • •To become a center of excellence available to the national university and research network and to MIUR, with the aim of spreading the culture of information and communication technology;

• •

  Developing research programs aiming at a more effective and innovative usage of information and communication technology, in collaboration with other organizations and enterprises;

In the field of virtual newspaper and periodical library, CASPUR allows many users (mainly coming from academic Italian institutions) to access to over 5200 academic and scientific full-text periodicals and over 7.5 millions pdf articles (last update: January 2009).

Journals are available dating the nineties; they cover all fields and are issued by different publishers and professional societies, including, for example, the American Chemical Society, Blackwell Publishing, Elsevier Science, Institute of Physics Publishing, Kluwer Academic Publisher, Springer.

This service is accessible from a web site (periodici.caspur.it) and its main advantage consists in the possibility of allowing research (also personalized) in different fields (author, title, keyword or full-text words) within the entire series. In this way users can refer to a title list arranged by publisher, class or alphabetical order, made possible by an homogeneous interface based on web-usability criteria.

Users access to the service and to the research function through a web client; this access is restricted to authorized Institutes and Universities through a procedure that checks the IP address or by considering a username and a password, which would allow the access to the virtual library from anywhere.

The virtual library service is based on Science Server software, and supplied by three Linux servers, indistinguishable by the final user. UltraATA disk strips (on 2 Gbps fiberchannel interface) form the disk space on which software, metadata and the indexes’ database are installed, for a total of 14 TB. Of these, 8 TB are dedicated to the online system, and the others are a copy of it, necessary to the whole system data backup

The idea of this study is to describe the behavior of the users by considering and analyzing their traces stored into the web server log file. The analysis of such logs can provide an insight about searching behavior on digital library and about Information Retrieval. It has to be noticed that the first in-depth studies on query logs date back to the late 1990s; see, for example, Jansen (1998, 2000) and Spink (2001). But, for what concerns the use of these files in a digital libraries context, there are less studies; see Wolfram (2002).

Materials And Methods

Data were collected by considering the web logs coming from those users that accessed to the digital library using a username and password; this facilitates the need to identify all the distinct search sessions (see further in the next paragraph).