International Business Machines Corporation

Donald H Randall, Jr

Nadiya Kochura

Timothy B Snow

Derek M Reedy

Vinodkumar Raghavan

Frank Robert Marullo

Frank Anthony Nuzzi

Charles Vaughn Rankin

Michael John Weisskopf

VinodKumar Raghavan Nair

Michael Vinh Vu

According to an aspect, software test failures are analyzed using natural language processing (NLP) and machine learning. A failure is detected during a code build associated with a software product. Each change set since a last successful code build associated with the software product is identified and analyzed using NLP to extract change set features. A software defect origin model is applied to the extracted features in each analyzed change set to detect an origin of the failure.

Analyzing software test failures using natural language processing and machine learning

According to an aspect, software test failures are analyzed using natural language processing (NLP) and machine learning. A failure is detected during a code build associated with a software product. Each change set since a last successful code build associated with the software product is identified and analyzed using NLP to extract change set features. A software defect origin model is applied to the extracted features in each analyzed change set to detect an origin of the failure. The software defect origin model includes a mathematical description of patterns learned from previously detected failures and their corresponding features.

A computer-implemented method is provided for determining quality metrics for a question set. In an implementation, a test question set model may be produced based upon calculated quality metrics of a test question set with respect to a test corpus, and including features representing quality metrics. The test question set model may be compared to a baseline question set model based on a distance calculated between one or more projected model features of the baseline question set model and one or more runtime model features of the test question set model. Contents of the test question set may be adjusted based upon the calculated distance.

System and method for determining quality metrics for a question set

A mechanism is provided in a data processing system comprising at least one processor and at least one memory, the at least one memory comprising instructions executed by the at least one processor to cause the at least one processor to implement a user interface automation tool. The user interface automation tool executes a script to perform automation functions on user interface controls in a user interface of an application. Responsive to automation of a given user interface control failing, the user interface automation tool identifies a candidate user interface control that is the same as a user interface control expected in the script using a machine learning model. The user interface automation tool corrects the script to refer to the candidate user interface control to form a corrected script. The user interface automation tool performs a user interface function on the candidate user interface control according the corrected script.

Automatically correcting GUI automation using machine learning

Releasing pending messages is provided. An electronic message is received via a network from a sending client device. The electronic message includes a request to place the electronic message in a pending file and an authorization to release data. The electronic message is placed in the pending file and the authorization to release data is stored in a storage device. An authorization is received to release the electronic message from the pending file. Then, it is determined whether the authorization to release the electronic message matches the stored authorization to release data. In response to determining that a match is found, the electronic message is released from the pending file. Then, the released electronic message is sent to a receiving client device via the network.

Sending and releasing pending messages

The present invention provides a method for process-specific exchange of data between machines in a computer network. Each machine runs a plurality of processes. The method includes providing a plurality of clipboards for at least one machine in the computer network, where each clipboard of the plurality of clipboards corresponds to one of the processes running on the at least one machine; and storing data concerning a process of the plurality of processes onto its corresponding clipboard. Because the method of the present invention is process-based, machines may obtain data concerning certain processes on certain machines without the problem of the latest data on a process being overwritten by another process. Since the clipboards of the present invention may reside on a third party, if one or more of the machines in the network fail or are otherwise unavailable, the data on their clipboards are not lost.

Method for process-specific exchange of data between machines in a network

Automated client-based web application URL link extraction tool for use in testing and verification of internet web servers and associated applications executing thereon

A computerized subsystem starts multiple user-defined threads, each of which is a virtual browser for a web site server application under test and executing on a web server. Each individual virtual browser requests and posts data to the server and corresponding server application under test at a rapid rate. Actual returned data is not saved but time of transaction and whether page is found and returned is logged. The server and port to test are user-specified in an input data file or can alternatively, by command line option. A repeat option causes the subsystem to loop through the input file a specified number of times, hours, and an indefinite repeat. User-specified sleep values slow down requests to simulate actual users. A wait option creates a semaphore causing all threads to wait for the semaphore which appear only after all threads are created, whereupon such threads begin execution simultaneously to maximize stress on the server and application. The wait option may be user-disabled whereby threads begin execution as soon as created for a staggered stressing of the web site and application. The input data file has a common format shared both by the web server stress tool and a web server application verification tool, multiple instances of which comprise the web stress tool.

Automated client-based web server stress tool simulating simultaneous

An Internet website virtual browser application automatically exercises and verifies web server applications and scripts by simulating a web browser to request, capture, store, and verify data returned from web servers, discarding data not critical to testing, and saving and reusing retained data for subsequent transactions. Input and links are accepted from a GUI edit field or input data file, with posts and gets supported to return pages and other requested data, saved in user-specified files for verification. GUI edit field options may override server/port definitions without changing input data files and provides for user-specified termination after a specified number of errors, specifying number of retries and level-setting of messaging for log files and screens. A log file stores the request and byte count total bytes transferred, time of transaction, throughput or transfer time, and result of compares. The verify option automatically compares previously received and current request data. A smart pass/fail status derives a total pass/fail count for all requested pages. Input data may be looped for reliability testing. Critical requests may be labelled and required to be successful; and sleep values between requests are user-specifiable to simulate actual users and test session timeouts.

Growing community of inventors

Austin, TX, United States of America

Donald H Randall, Jr