International Business Machines Corporation

Gaurav Tewari

Reiner Kraft

Michael Lawrence Emens

Daniel Alexander Ford

Shang-Hua Teng

Peter Chi-Sing Yim

The present invention discloses a method for dynamically generating an abstract from a search result abstract retrieved from a search engine. Preferably embodied as a software program, the present invention allows a user to 'zoom' in and out of the abstract thereby retrieving greater or lesser details of the abstract to further determine whether the resource is useful or not. Groups of words within the search result abstract are searched for indexable words or key words useful to the user. The indexable words are then displayed to the user. Advantageously, the process of the present invention is performed entirely on the user side without having to reset the abstract level at the server side with a new query.

Technique and method which dynamically generates various depth levels of abstracts for text based data sources on demand

A method, system and program storage device are provided for interjecting messages into a real-time isochronous discourse (such as a telephone call) between a plurality of callers wherein the method, system and device accesses the call between the callers, monitors the discourse to determine if the discourse relates to a system message desired to be communicated to the caller, chooses and communicates the system message to the callers when the discourse is determined to be relevant to the system message and continues the above accessing, monitoring and communicating steps until the call being accessed between the callers is terminated. The interjected messages are primarily related to advertising and can be introduced to one or more callers, or, separate messages in different languages can be introduced to one or more callers depending on the caller's accessing the telephone. Caller subscribers of the system will typically have the cost of the call partly or totally subsidized by the method and system.

System and technique for dynamically interjecting live advertisements in the context of real-time isochronous (telephone-model) discourse

A system for accepting a specification of a time interval during which a download is to be performed on a delayed basis is described. The system can also select a time within the specified period. The selection can be randomized with the aim of distributing the load placed on computer and communication resources, or the selection may be based on testing a number of measures of computer and communication resource availability, and selecting a time of low utilization. A fuzzy measure of resource availability can be obtained, e.g., as a weighted sum of multiple measures, or each measure can be tested against a limit separately. According to another aspect of the invention a system and computer readable medium is disclosed for carrying out the above method.

Software and method for controlling the timing of delayed downloads

A method of monitoring real-time communication on a computer network between at least two client computers connected by the network comprising first providing a database of keywords, with each of the keywords linked to a predefined rating. The method then includes monitoring real-time communication on a computer network between at least two client computers connected by the network; detecting the keywords in the real-time communication; and determining for the real-time communication a rating level based upon the predefined rating of the keywords. The keyword may consist of text, audio, video and graphical communication, for example. Preferably, the rating level of the real-time communication is conveyed to at least one of the client computers. More preferably, at least one additional client computer receives the real-time communication, and the rating level of the real-time communication is conveyed to the at least one additional client computer. Determination of the rating level for the real-time communication may occur simultaneously with the real-time communication, and may be based on evaluation of individual ratings of a plurality of different keywords. The method may further include terminating the real-time communication of at least one of the client computers based upon the rating level.

Platform for internet based real-time communication content selection

A method in a computer network for identifying users with similar interests is described. The method includes accepting a first query statement from a first user and storing a first item of information related to the first query. The method further includes accepting a second query from a second user and storing a second item of information related to the second query. The method further includes computing a measure of similarity of the first query and the second query by using the first item of information and the second item of information. A system and computer readable medium for carrying out the above method is also described.

Internet based method for facilitating networking among persons with similar interests and for facilitating collaborative searching for information

A method of selecting a mirror server for client-host interaction over a network provides automatic, user-transparent, selection of a particular mirror server from among multiple mirror servers that replicate a host server providing web content over the Internet. The method selects the mirror server that has the best response time for a given client computer requesting web content under changing web traffic and mirror server load conditions. The host server creates and maintains a list of addresses of available mirror servers, preferably ranked according to the load on the mirror servers. In response to an initial information request from the client computer, the host server sends the mirror server list to the client computer along with the requested web content. The client computer creates multiple mirror server requests, corresponding to the mirror servers on the address list, sends the requests to the mirror servers, then compares the round trip response times to determine the optimum mirror server site from which to continue subsequent client-host interaction. The invention also includes methods and apparatus, preferably in the form of computer program products, which are operable at the host server, to create, maintain and send the address list, and other methods and apparatus operable at the client computer, to receive the list, query the mirror servers and select the optimum mirror server for further interaction.

Method of automatically selecting a mirror server for web-based client-host interaction

A method and apparatus for automatically gathering, summarizing, and indexing real-time information derived from real-time communication on the Internet, such as Internet chat sessions, or any other comparable form of real-time communication on the Internet. The present invention provides live information selection and analysis and provides basic functionality of crawling, indexing and summarizing chat room data, as well as generating messages, such as advertisements, responsive to the subject matter of the communication.

System and technique for dynamic information gathering and targeted advertising in a web based model using a live information selection and analysis tool

A computer program product is provided as a system for use with a search engine and a search engine repository to dynamically associate keywords within a search result set encountered in an abstract of a search result item with a domain-specific query. In the system, a local query database stores the domain-specific query, and a synchronization unit synchronizes the search engine repository and the local query database. A query builder builds a search query from a query template using the search engine repository. A summary marker incorporates the search query with the keyword in the abstract of the search result item. A keyword detector generates a list of keywords included in a domain-specific dictionary. A search result calibration manager calibrates the number of the query result items. A search result item buffer receives a request for processing an abstract metadata item from the synchronization unit. The request includes a Universal Resource Locator (URL) and a corresponding abstract, wherein the query builder uses a (URL, keyword) pair to build the domain-specific query from the query template. The summary marker updates the abstract corresponding to the URL in the search engine repository, and marks and inserts the domain-specific query for all occurrences of the keyword.

Growing community of inventors

Cambridge, MA, United States of America

Gaurav Tewari