Abstract: A spell checker based on the noisy channel model has a source model and an error model. The source model determines how likely a word w in a dictionary is to have been generated. The error model determines how likely the word w was to have been incorrectly entered as the string s (e.g., mistyped or incorrectly interpreted by a speech recognition system) according to the probabilities of string-to-string edits. The string-to-string edits allow conversion of one arbitrary length character sequence to another arbitrary length character sequence.

Abstract: Contents to be embedded in a page template or layouts can be dynamically changed without editing an HTML file or correcting a server program such as CGI. A page template is defined with a servlet that specifies a display area and a display attribute for controlling displays as parameters. When the Servlet is recognized at the web server side, a plurality of part beans that are stored by being related to display-area specifying information are specified. The information for linking image data for forming an object to be embedded in the display area and link-destination-URL information are set to the part beans as properties. Moreover, display-condition information such as schedule information or the like is related to the part beans.

Abstract: A method and apparatus for synchronizing audio data to text data in which the text is manually or automatically divided into an arbitrary number of sections and a text section is selected for display and presented to a narrator for audio input. A human narrator reads the displayed selected section of text into a computer to form an audio data portion for the selected section that is linked to the text by entering synchronization marks to the computer. The sections of text may be selected for reading and audio input in any order and by any number of narrators. Each recorded audio portion is converted to an audio data file and the audio data files are re-sequenced to conform to the original order of the text based on a playback control file. A plurality of narrators may contribute audio data from separate workstations of a network, or workstations which are connected to a central server over a communications network, such as the Internet.

Abstract: A number of items of data from a data source (12) can be processed, and then supplied to a data destination (16-17). The data can include image data, text data, numeric data or some other type of data, or a combination of these types of data. The processing of the data is controlled by a project definition (14, 71, 101), which includes a plurality of modules selected from a variety of available modules. The modules have input and output ports, and the project definition includes binding information which relates input ports to output ports. The binding information can include a conditional binding (128), which supplies data to an input port from a selected one of two or more output ports, in dependence on a condition. A separate feature is that, if data obtained in a given format from an output port is being supplied to an input port which expects data in a different format, the data is automatically converted to the format expected by the input port.

Abstract: The system, method and program of this invention enhances a Web browser by enabling a user to select hyperlinks of interest while reading a first document. The selected hyperlinks are not displayed by the Web browser until indicated to do so by a user. In the meantime, the Web browser stores the list of selected hyperlinks in a list. When a user finishes reading the first document, and now desires to view the selected hyperlinked documents, the user selects a “finished reading” button. Upon the selection of the button, the Web browser displays a next hyperlinked document according to the order of selected hyperlinks in the list.

Abstract: An application server, configured for executing an application defined by XML documents for delivering voice portal services, accesses subscriber attributes from a standardized information database server (such as LDAP). The application server, upon receiving a request from a proxy browser serving a user, accesses the standardized database server to obtain attribute information for responding to the application request. The application server sends a request to a content server for media content based on the HTTP request and the attribute information. The application server generates an HTML page, for execution of the prescribed voice portal service application having XML tags configured for controlling delivery of the media content in an audible format, based on the HTTP request.

Abstract: One embodiment of the present invention provides a system that generates an index to facilitate searching through text within a document based upon an index stylesheet associated with the document. The system operates by receiving a document to be indexed and then parses the document to produce a parsed document. The system also retrieves instructions for creating the index for the document from an index stylesheet associated with the document. The system creates the index for the document by transforming the parsed document in a manner that is specified by the instructions retrieved from the index stylesheet. In one embodiment of the present invention, retrieving the index stylesheet involves retrieving the index stylesheet across a network from a remote address.

Abstract: A self-modifying data flow architecture for computer-readable structures, such as markup language, is modeled as a network of interconnected processing elements, each having a data input and a transformation input. Each processing element generates output by applying the transformation input to the data input. The output of one processing element may be provided as either a data input or a transformation input to another processing element. The resulting architecture provides a network of interconnected processing elements which are modified dynamically depending on the data flowing through the overall process.