International Business Machines Corporation

Ajoy Acharyya

John Michael Ganci, Jr

James Edward Bostick

Craig M Trim

Shuvamoy Saha

Souvik Majumdar

Arijit Roy

Disclosed embodiments provide techniques for automatic software defect correction of a computer program. Computer program log files are scanned to identify runtime errors, corresponding to software defects. The software defects are analyzed to determine an error type, and identify the source file/code that caused the error. A solution template repository is searched for a solution template corresponding to the identified error type. If a solution template is found, the source code is checked out from the identified source repository. The template is applied to the 'original' checked out source file to create a new source file with the fix, which is then uploaded back to the repository. A new software distribution is automatically built with the new source file, and the new software distribution is automatically deployed to the devices that experienced the error. Thus, defects can be automatically detected, repaired, and deployed without human intervention.

Automatic software defect repair

A computer processor determines a schema that enables splitting of one or more elements of an XML file. The computer processor determines an XML file as a split candidate, based on one or more attributes of the one or more elements of the XML file. The computer processor splits the XML file at run-time into a plurality of subsets of the XML file, based on the one or more attributes of the one or more elements of the XML file, and the computer processor distributes the plurality of subsets of the XML file to a plurality of computing nodes of a computer processing system.

Processing large XML files by splitting and hierarchical ordering

A method, system, and computer program product for selectively disseminating content in social media are provided in the illustrative embodiments. Prior to a dissemination of the content, the content being prepared for the dissemination is detected. Using Natural Language Processing (NLP), the content is analyzed to identify a theme and a subject matter of the content. A set of data comprising data related to a set of social connections of a user in the social media is analyzed. A second theme resulting from analyzing a first data in the set of data is compared with the theme, the first data corresponding to a first social connection in the set of social connections. When the second theme matches the theme within a threshold, the first social connection is added to a selected sharing group. The content is disseminated to the selected sharing group.

Growing community of inventors

Kolkata, India

Ajoy Acharyya