Abstract: A document is stored in and retrieved from a repository being indexed by a search engine by applying to each document a first classification based on a profile of a user who is likely to access the repository. A query formulated by the user to the search engine results in the engine producing a list of references of documents. The list of references is reduced by eliminating from the lists the documents and references which do not correspond to the user's profile. A second classification level, representative of the particular nature of the problem to which the document refers, is performed. Automatic tagging and labeling mechanisms improve the repository partitioning.

Abstract: A process for accelerating search in a repository of resources indexed by a search engine. Each resource is classified in accordance with a least a first classification level. The process comprises the steps of: monitoring the access to said repository by the users and creating a shortcut index of resources which are flagged by the users in order to develop experience from the use of the repository; processing each new query entered by one user by a first preliminary searching within said shortcut index, followed by a standard search within the whole repository.

Abstract: A web server is described having a control arrangement comprising a state machine that is arranged to process an event initiated at a browser to transition from one state to another and a mechanism for generating, in response to entry of the states, corresponding web pages for transmission to the browser. The state machine comprises at least one superstate that comprises first and second states, the first state being selected by default on entry to the superstate, and at least one transition is defined from the superstate to the second state, the transition being associated with an action that provides to the browser information that is not destined to viewed in the browser. A web page generated in response to entry of the first state includes at least a link corresponding to the transition and the arrangement is such that entry into the second state does not result in the generation of a web page for transmission to the browser.

Abstract: Process for measuring the performance of a information retrieval system comprising a repository of documents or files associated with a search engine; said process involving the steps of:

Abstract: A document is stored in and retrieved from a repository being indexed by a search engine by applying to each document a first classification based on a profile of a user who is likely to access the repository. A query formulated by the user to the search engine results in the engine producing a list of references of documents. The list of references is reduced by eliminating from the lists the documents and references which do not correspond to the user's profile. A second classification level, representative of the particular nature of the problem to which the document refers, is performed. Automatic tagging and labeling mechanisms improve the repository partitioning.

Abstract: A process for accelerating search in a repository of resources indexed by a search engine. Each resource is classified in accordance with a least a first classification level. The process comprises the steps of: monitoring the access to said repository by the users and creating a shortcut index of resources which are flagged by the users in order to develop experience from the use of the repository; processing each new query entered by one user by a first preliminary searching within said shortcut index, followed by a standard search within the whole repository.