http://useit.com. the site of one of the leading specialists in Web ergonomics, Jacob Nielsen, contains many interesting considerations concerning the Web, Wap and the development of communications objects. In brief, the success of the Internet and the Web from the mid 1990s onwards led observers to believe that it was possible to design almost anything in the field of communications objects, in particular in terms of information and document access. Fairly soon, however, it became clear that more than half of the information searches performed on the Web resulted in failure. Furthermore, Web and Wap access via communications objects such as mobile telephones or PDAs, which were initially considered to be a development or extension of the Web, have not proved themselves to be particularly usable and are employed only very infrequently. A number of hypotheses have been considered on the basis of this observation. It was thought that technical advance, in particular in terms of data rates would solve some or all of the problems relating to the development of any given tool. Other analysts suggested that users' skills would develop, as they always do, and thus solve the utilisation problem. In brief, the development of usability was expected to boost utilisation.
In this chapter we intend to defend a different point of view. We may consider that, considering among other things its usability characteristics, any given communications object is specifically useful for a particular information-seeking task but not for others. A description and a categorisation of the information-seeking tasks and the establishment of a relation between this categorisation and the communications objects which do or do not permit the implementation of each of these tasks would allow designers to choose the most appropriate tools in the light of the services that they want to develop. According to Tricot and Nanard [TRICOT, 1998] the description of an information task should take account of:
The user 's representation : the representation that users construct of the task and their level of expertise in the field in question together with their skill in using the tool or the information service;
The implementation of the goal : the address and the number of targets in the system, the procedures to be used in order to access these targets, the general structure of the system and the interface;
The context of the activity : the reason why a subject chooses to use a particular system in order to search for certain information (learning, document design, problem solving etc.).
We have defined four objective variables which are independent of the user or the topic in question, and which make it possible to characterise the implementation of the goal in an information-seeking task. These variables are the repetitiveness of the task, the level of explicitness of the targets, the location of the targets and the quantity of targets. We have started to analyse the hierarchical relations which may exist between these variables . This has enabled us to describe 12 information-seeking tasks and classify these tasks on the basis of the mean performances they result in during information searches on the Web.