Abstract: Data communications apparatus and methods employing a bidirectional low bandwidth channel and a unidirectional high bandwidth channel. A PC is coupled to an information source by a communications system which provides both a bidirectional low bandwidth channel between the PC and the information source and a high bandwidth channel in which the information source is the source and the PC is the sink. A component of the communications system termed the director responds to a message received in the low bandwidth channel by switching the information being sent from the information source to the PC to the high or low bandwidth channel as specified in the message. The message may come either from the PC or the information source. A graphical user interface at the PC provides the user with "buttons" to specify the bandwidth. The apparatus and methods may be advantageously employed to provide telecommuting services.

Abstract: Apparatus for providing displays in which the locations of nodes of a graph relative to each other are a function of the significance of a relationship between the nodes. The technique used to locate the nodes produces groups of nodes with significant relationships to each other which are set visually apart from the other nodes. The relationships between the nodes are defined by using a statistic to give weights to links between the nodes. Statistics may also be used to determine the size, shape, and color of the nodes and the color and width of the links. The display may also include histograms which show the numbers of links or nodes which have given values of the statistics and which relate the colors in which the nodes and links are displayed to values of the statistics. The histograms may further be used to interactively define masks for links and nodes, and masked links and nodes are not taken into account in making the display.

Abstract: A system which is a combination of editor, window system, shell, and user interface and which provides a novel environment for the construction of textual applications such as browsers, debuggers, mailers, and so on. The system combines an extremely lean user interface with some automatic heuristics and defaults to achieve significant effects with minimal mouse and keyboard activity. The user interface is driven by a file-oriented programming interface that may be controlled from programs or even shell scripts. By taking care of user interface issues in a central utility, the system further simplifies the job of programming applications that make use of a bitmap display and mouse.

Abstract: A backup file system which can be used without modification of application programs, the operating system, or the hardware. The backup file system is implemented by means of a dynamically-linkable replacement library and user-level processes on a primary computer system and a backup computer system. The dynamically-linkable replacement library has the same interfaces as a standard dynamically-linkable library of file operations. The functions in the replacement library perform the same file operations as their counterparts in the standard library; in addition, they send messages specifying the just-performed operation to a user-level process on the backup computer system. The user-level process executes the operation specified in the message on a backup file in the backup computer system. The files to be backed up are specified by identifying subtrees in the namespace of the primary file system; the subtrees make up a user-defined namespace.

Abstract: A library of C++ classes for use in writing data visualization programs. The library embodies a general design principle for class libraries: that the classes are partitioned into entity classes and functionality classes. Entity classes in the library specify properties of areas in a display and the functionality classes specify functionalities which the areas may have, for example, being responsive to the mouse or the keyboard or being capable of executing drawing commands. The classes are narrowly defined, and when the C++ programmer specifies a class for an object, he or she can give the object exactly the desired properties by selecting from among the area classes and the functionality classes. Because properties are acquired by inheritance, code generated using the libraries is shorter and faster than code generated using conventional class libraries.

Abstract: Apparatus and methods for specifying contexts for machine-executable instructions. Modern graphical user interface systems employ the callback programming style. In this style, a system event handler responds to an event by executing application-level callback code and providing event information concerning the event as part of the context of the execution. The technique disclosed herein uses callback information names to specify the relationship between the event information and the execution context. The callback information names are defined globally, but represent the event information for a single execution of the callback code. Using the callback information names, it is possible to define the callback code in the function which adds the callback code to the graphical user interface system.

Abstract: Apparatus and methods for determining whether a word/sense pair is proper for a context. Wide contexts (100 words) are employed for both training and testing, and testing is done by adding the weights of vocabulary words from the context. The weights are determined by Bayesian techniques which interpolate between the probability of occurrence of a vocabulary word in a conditional sample of the training text and the probability of its occurrence in the entire training text. A further improvement in testing takes advantage of the fact that a word is generally used in only a single sense in a single discourse. Also disclosed are automated training techniques including training on bilingual bodies of text and training using categories from Roget's Thesaurus.

Abstract: Apparatus and methods for reducing the amount of computation required to determine the consequences of a side effect resulting from a transaction on a relation in a data base system. The apparatus and methods use equational reasoning based on a bag algebra representation of a class of transaction and the relation to produce a pre-expression which describes the side effect. The pre-expression is included in the data base system's data dictionary, and when a transaction of the class is received in the data base system, the pre-expression is used to determine the side effect. The pre-expression is formed with a strongly-minimal pair of delete and insert bags. The apparatus and methods may be used to maintain materialized views, to check for violation of constraints on the data base, and to activate triggers, monitors, and active queries.

Abstract: A compact representation of a tree data structure and techniques for navigating the compact representation. The compact representation is a list. Each element of the list represents a node of the tree and the list is organized according to a preorder traversal of the tree. Each list element contains only the index of a data dictionary entry for the kind of item represented by the node corresponding to the list element. The navigation techniques permit the location of the list element for the sibling of the node corresponding to the given list element, the location of the list element for that node's parent, and in the case of a parent node, the location of the list element for any child of the parent. The navigation techniques work by finding the list elements for subtrees. The subtrees are found by techniques based on the fact that the number of children of all of the nodes in a subtree minus the number of nodes in the subtree always equals -1.

Abstract: A technique which may be used in a transaction which updates a data set when the data set is subject to constraints and includes operations which may abort. The technique produces an update constraint which may be tested before the data set is altered to determine whether the constraints will be satisfied if the data set is altered. The update constraint of the technique indicates whether the constraints would be satisfied if the transaction could not abort. The update constraint is checked at the application level, while the constraints which cause an operation to abort are checked at the system level. An application of the update constraint to a relational data base embedded in a telephone switch is disclosed.

Abstract: Apparatus for developing and verifying systems. The disclosed apparatus employs a computationally-tractable technique for verifying whether a system made up of a set of processes, each of which has at least one delay constraint associated with it, satisfies a given temporal property. The technique deals with the verification as a language inclusion problem, i.e., it represents both the set of processes and the temporal property as automata and determines whether there is a restriction of the set of processes such that the language of the automaton representing the restricted set of processes is included in the language of the automaton representing the temporal property.

Abstract: Apparatus and methods for incorporating a portion (including all) of a first name space into a second name space. An agent operating in the first name space receives first operation specifiers which use names as they are used in the second name space from an entity operating in the first name space and translates the first operation specifiers into second operation specifiers which specify the same operations and use the names as they are used in the first name space. The apparatus and methods are implemented in the Plan 9 operating system. The first and second name spaces are organized as trees and belong to a first process operating on a first processor and a second process operating on a second processor. The first operation specifiers are Plan 9 file protocols which the first process receives from the second process. The second operation specifiers are Plan 9 file system calls made by the first process.

Abstract: A file system which has component file systems including a primary file system which is react/write and a number of dump file systems which are read only. Each dump file system is created from the primary file system by means of a dump operation and conserves the state of the primary file system at the time the dump operation was performed. Component file systems share read only storage elements with older component file systems. The file system is implemented on a system including a file server, a magnetic disk mass storage device, and an optical write once-read many (WORM) disk. The magnetic disk mass storage device contains the read/write storage elements of the primary file system and encached read only storage elements from the WORM disk. Space is reserved on the unwritten portion of the WORM disk for the read-write storage elements of the primary file system.

Abstract: Techniques for automatically laying out directed graphs with clusters of nodes and free nodes. The techniques produce pictorial representations of the graphs in which a node in a cluster and a free node may belong to the same rank in the graph and nodes belonging to more than one cluster may belong to the same rank. The techniques have been added to the well-known DOT system for laying out directed graphs, and include a recursive technique for ranking nodes of the graph, a technique for ordering nodes within ranks which collapses subclusters into skeletons when the nodes within a cluster are ordered, and a technique for positioning the nodes after they have been ranked and ordered which obtains the x coordinates for the nodes by producing an auxiliary graph which is in effect the original graph "turned on its side" and assigning ranks to the nodes in the auxiliary graph.

Abstract: The use of nonmonotonic or "common sense" reasoning has often been advocated to speed up reasoning systems. The difficulty has always been that nonmonotonic reasoning requires consistency checks, which are even more computationally intractable than the monotonic reasoning normally used in reasoning systems. Disclosed herein are methods and apparatus for making consistency checks tractable by performing them over a limited context using fast sufficient consistency checks. The results are approximate, but approximate results are expected in nonmonotonic reasoning. Included in the disclosure are techniques for selecting the limited context and making the fast sufficient consistency checks and a default reasoner which employs the techniques to check the consistency of a default conclusion with a limited context.

Abstract: A processor specially adapted for use as a coprocessor. The processor is implemented as a microprocessor. The adaptations include the following: The microprocessor has a master-slave pin which receives an input which determines whether the microprocessor operates as a bus master or a bus slave. Certain output pins have three-state bus drivers which employ feedback to ensure that a signal on a line being even by the driver has gone inactive before the driver is turned off. Instructions executed by the microprocessor permit specification of portions of the internal registers as sources and destinations and specification of the size of an ALU operation, permitting easy operation on data ranging from bytes through 24-bit pointers. Instructions are executed in an instruction pipeline and a separate I/O instruction pipeline.

Abstract: A processor specially adapted for use as a coprocessor. The processor is implemented as a microprocessor. The adaptations include the following: The microprocessor has a master-slave pin which receives an input which determines whether the microprocessor operates as a bus master or a bus slave. Certain output pins have three-state bus drivers which employ feedback to ensure that a signal on a line being driven by the driver has gone inactive before the driver is turned off. Instructions executed by the microprocessor permit specification of portions of the internal registers as sources and destinations and specification of the size of an ALU operation, permitting easy operation on data ranging from bytes through 24-bit pointers. Instructions are executed in an instruction pipeline and a separate I/O instruction pipeline.

Abstract: Apparatus and methods used in a network for translating names into network addresses. Source nodes in the network receive names and have name caches for translating the names; destination nodes have local name tables containing correspondences between names and the network addresses served by the destination node. Them is further a name server accessible to the nodes which contains current correspondences between names and network addresses. A name is translated by first presenting it to the name cache in a source node; if there is no entry for the name in the cache, name cache maintenance code in the source node obtains information about the name from the name server; once there is an entry, the name is translated and a call set up message containing the name and the network address is sent to the destination node, which checks whether the name and network address are its local name table. If they are, the destination node returns a call complete message to the source node.

Abstract: Apparatus and methods for producing segmentations of images which contain text. The general approach is to locate a first set of components which contain text characters, locate a second set of components which do not contain text characters, sort the second set using a characteristic shape, make a cover set from the sorted second set, and use the cover set to locate the portions of the image which contain text. An embodiment is disclosed in which the method is applied to texts employing the Manhattan layout to locate columns of text. In the embodiment, parts of the image which do not contain text are located by constructing maximum empty rectangles which do not contain characters. Based on the observation that columns in Manhattan layouts are separated by rectangles with a high aspect ratio, the maximum empty rectangles are sorted in a manner which favors such rectangles to produce the cover set.

Abstract: Memories which have new benign failure modes. The new failure modes are the omission-crash mode and the eventual-crash mode. Memories having either of these modes fail more benignly than memories having the omission failure mode, but unlike memories having the crash failure mode, memories with the novel failure modes may be used to construct gracefully-degrading fault tolerant objects. Implementations of memories with the new failure modes are disclosed, along with implementations of fault-tolerant objects made from memories with the new failure modes.