Abstract: A method and apparatus for controlling a hardware merge sorter to sort N.gtoreq.2.sup.n number of records, the hardware merge sorter including a plurality of local processors that are cascaded in series, a plurality of local memories and a plurality of selectors. Each selector is coupled to and between one of the plurality of local memories and a respective local processor and couples the local memory to the respective processor during a first sorting step. In addition, the plurality of local memories are coupled to one local processor, of the plurality of local processors, during a second sorting step. With this arrangement, the hardware merge sorter can sort a plurality N.ltoreq.2.sup.n+1 number of records. In one particular embodiment of the present invention, a plurality of hardware merge sorter cells are cascaded in series. With this arrangement, the hardware merge sorter can sort a plurality N.ltoreq.2.sup.n+2 number of records.

Abstract: The improved Quicksort method of the present invention utilizes two pointers initialized at opposite ends of the array or partition to be sorted and an initial partition value Pvalue located at the center of the array or partition. The value at each of the end pointers is then compared to Pvalue. Sorting is accomplished by recursing the partition process for the two array segments bounded on one side by the final P valve location. This method prevents excessive recursions, and allows the identical array case to recurse to the ideal minimum: Log2(N). Also, by relaxing the offsider criteria to include elements equal to Pvalue, the present invention presents arrays of two valves or a very small range of valves from recursing excessively. Further, the sorting method of the present invention may test the final position of the initial Pvalue to determine whether it is in the center (75%-95%) portion of the array or subarray being positioned.

Abstract: A computer database method wherein the data is organized into atomic data sets and transactions are separated into projections which operate on only one atomic data set. Multiple transactions can thereby access the same atomic data set using a plurality of projections. A projection timestamp is generated by the system initial access to the atomic data set by the projection. Read and write timestamps are maintained separately by the atomic data set. Upon each read and write command, the issuing projection timestamp is compared against the appropriate atomic data set read or write timestamp. The operation succeeds if the compared timestamps do not indicate a conflict. Upon failure the projection is restarted with a new timestamp.

Abstract: A method and apparatus for transferring control between application programs. A messaging means is provided which allows a first application program to indicate to the messaging means that a second application program should assume control. The messaging means receives the message and performs an orderly shutdown of the first application program and messages the second application program that it should commence operation. Upon valid and proper operation of the second application program, the first application program is caused to be suspended, and the second application program is invoked.

Abstract: An Electronic Computer Aided Design System provides for concurrent operation of a plurality of design tools which share a common design dataset. Changes made by one program to the design dataset are immediately updated and are automatically reflected in the displayed outputs of the other design tools. A tool manager program allows rule-based automation of the entire system.

Abstract: A method and system are disclosed for formatting address strings into a recognizable sequence of token types for database processing. The system includes a token rule processor and a token sequence processor. The method begins with the step of assigning token types to components of an address string to form a sequence of token types. The method next includes the step of determining whether or not the sequence of token types is contained in an adjustable predetermined rule table. The method further includes the step of processing the sequence of token types into a recognizable sequence format if the sequence of token types is contained in the rule table. Finally, the method concludes with the step of processing the sequence of token types into a recognizable sequence format in accordance with a predetermined interpretation procedure if the sequence of token types is not contained in the rule table.

Abstract: A computer database method using data organized data into atomic data sets and transactions separated into a plurality of statements called projections which operate on only one said atomic data set. Statements such as read commands, write commands, assignment commands, transaction delimiting commands, and projection delimiting commands are executed, and a depends on graph is created during execution of the statements. Projection recovery data is stored during operation. When a projection requests commit the depends on graph is analyzed to determine if the projection is depends on another projection in the transaction. The projection is committed if it is independent or delayed until commit of the projection on which it depends. Likewise, the system determines projections dependent on the aborting projection and aborts the projections together if a transaction error occurs. A replay method is presented for restarting aborted projections.

Abstract: A method and apparatus for performing a document information search to uncover specified text data containing a given search subject key word from a group of document text data stored in a memory. In the document information search method, two stages of presearch are carried out to perform the document search with respect to a desired subject key word. In a first stage of presearch, a character component table is generated in which the existence of character codes for every document is set forth with respect to all the character codes contained in the group of document text data of stored documents. The character component table is searched for all the character codes comprising a designated search subject key word to thereby extract all documents containing all the character codes comprising the search subject key word. Further, in the presearch step, all texts without the possibility of containing the search subject key word are eliminated.

Abstract: A retrieval interface system which allows a user to view a retrieval result as a percentage of a total number of files, and to further classify the retrieval result into groups based upon attributes of the files. The retrieval interface system includes a classifying section which classifies the contents of a database based on a classification method, a display management section, a classification method storage section, a control section, a history information storage section which retains history information of retrieval conditions, and a quantity display conversion section which converts display elements into a form from which the user can determine the number of elements in each group.

Abstract: A multi-media terminal apparatus controls switches to perform communications, so that processing units of its own terminal which controls various multi-media processing functions are connected to other multi-media terminal apparatuses via a transmission line having a plurality of channels in accordance with connection request instructions, or this connection state and connections among the processing units are simultaneously attained.

Abstract: A method and apparatus for conversion of a document from one format to another format. The invention inputs a binary file and determines the cleartext name for each ASN.1 tag in the binary file. An element table is used to store the features of each element and the sub-elements which are permitted to appear under each element are stored in sub-element linked list data structures referred to by the element table. It is possible for the binary encoding to be in an indefinite or definite length format and an element stack is used to keep track of the length of the binary file. The element stack is also used to keep track of the elements which are being processed in the hierarchy of the document. After a binary element is converted, the information pertaining to the cleartext representation of the binary element is stored in a double linked list data structure which refers to a previous element and a next element.

Abstract: A method and system for converting a document from a clear text representation to a binary representation. There is not a one to one correspondence between the two formats in SPDL. In order to determine what binary representation is needed for a clear text tag which can have a plurality of binary representations, it is necessary to actually process at least portions of the document for presentation (in contrast to just converting the document to a new format) in order to determine which binary representation of the tag is to be used. A stack is used to keep track of the binary elements as they are being converted. This allows for an efficient handling of the binary length encoding because the length of the binary representation of the different hierarchical levels of a document are not known until the respective hierarchical levels are encoded.

Abstract: In a database accessing system, access information concerning access to a database system is transferred using call control messages including user-to-user information within a control procedure defining the connection of a call on a signal channel. This results in a database accessing method that offers the improved use efficiencies of communication lines and the upgraded operating ratio of a database system, and allows users to enjoy reduced communication charges.

Abstract: The present invention provides an expert system for suggesting and providing graphic design selections based on a user's desired result. The invention provides a plurality of descriptors to define the characteristics of an output page or document. A user can "scale" or "weight" the descriptors, as desired. The present invention utilizes the weighted descriptors to select a number of solutions from a predefined database of possible solutions. The "fit" of each database solution to the user criteria is determined via an intelligent search. Initially, solutions that do not satisfy rejection constraints are identified and eliminated. These rejection constraints are binary constraints and may be the result of output device limitation, intended audience, environment of use, availability of color, etc. The elimination of solutions that do not meet the rejection constraints reduces the population of possible solutions, simplifying further researches and analysis.

Abstract: A data storage and retrieved system and process for use with structural items of data relying on associative indexes to connect related items of data and not to connect unrelated items of data. The associative indexes being formed from mathematical and/or boolean expressions for allowing know-how of a expert, stored in the system, to be extracted and put into context to a particular situation by a user inputting choices among alternatives and data into the system.

Abstract: A method and apparatus for increasing the speed of a search through leaf nodes of an inventive key index tree structure, known as a "Q-tree". The present invention identifies "decision-bits" within leaf node entries and uses these decision-bits to accelerate searches for a particular record or for a record adjacent to which a new record is to be inserted. A decision-bit is a particular type of distinction-bit having the smallest value and associated with a search key a greater value then the keys associated with other distinction-bits of the same value within a specified search range. Each decision-bit divides a search range into "left" and "right" parts. One of the two parts will constitute a "new" search range. A "quit-bit" having an ordinal number equal to the value of the decision-bit for the search range, is tested. If the quit-bit is "on", then the right, or greater value, part becomes the new search range. Otherwise, the left, or lesser value, part becomes the new search range.

Abstract: The present invention is directed to a technique extending the commands, and consequently, the capability of a system editor. Rather than extending the editor with existing editor commands, as in the use of macros, the present invention enables the user to extend the editor using new commands, called transforms. The transforms are written in a standard high level language such as ALGOL, C, COBOL, FORTRAN, or PASCAL. A transform is a new type of command that may be added to a system editor. The transform command is created, compiled, and then stored in a library. When the command is called, an interface in the system editor provides access to the transforms in the library. Transforms are designed so that, to the user, they operate like normal editor commands.

Abstract: Computerized tools for modeling database designs and specifying queries of the data contained therein. Once it is determined that an information system needs to be created, the Fact Compiler of the present invention is invoked to create it. After creating the information system, the user creates a fact-tree as a prelude to generating queries to the system. After creating the fact-tree, the user verifies that it is correct using the Tree Interpreter of the present invention. Once the fact tree has been verified, the Query Mapper of the present invention is used to generate information system queries.

Abstract: A database management system including an index with parallel read access for large amounts of keyed information. The index can be referenced for read-only access by any number of simultaneous processes in parallel with being updated by a single process where the read-only access does not involve any locking or waiting.

Abstract: Online prediction techniques based on data compression principles are employed to make predictions in restricted memory environments. Predictors have data structures in the form of trees that are paged and maintained in a cache on a least recently used replacement basis. A fast sequence of events strategy increments the counts for events at the current node of the predictor.