Browse Prior Art Database

ACD - search problem reports for other defects that are associated with a developer's change set Disclosure Number: IPCOM000235573D
Publication Date: 2014-Mar-10
Document File: 3 page(s) / 48K

Publishing Venue

The Prior Art Database


Disclosed is a system used in an Integrated Development Environment (IDE) that detects regression in functionality/fixes covered by previously existing work items, to enable efficient and complete fixes to problems.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 42% of the total text.

Page 01 of 3

ACD - search problem reports for other defects that are associated with a developer's change set

In times of hectic development, communication of bugs/fixes across a software development team may be poor, particularly if the team is large and/or geographically dispersed. One developer's code changes may fix (or break!) some additional defect/tasks in the work request system of which that developer is unaware. In an environment with poor communication, the Integrated Development Environment (IDE) might have five defects open, all having the same root cause of the same NullPointerException (NPE) error, but the testers may not realize or understand that one root cause is to blame for all defects. Testers might not even know enough to set up 'duplicated', 'related', etc. links.

Additionally, a developer can frequently commit code for bug fixes and feature requests, and then not remember the specific work request number(s) on which the

work is being done. The developer can use a keyword search, but that only works if the exact correct keyword is entered (i.e. not a synonym). Furthermore, work

requests can languish in a backlog when forgotten, and a developer might easily fix the related bug/issue and never associate it with that backlogged issue, leaving that issue in an open status.

The solution is a system to detect regression in functionality/fixes covered by previously existing work items. The key novelty is in discovering what work requests could be affected by a current change, beyond the work items in a developer's immediate range of knowledge.

The system enables developers, when delivering a code block, to take advantage of all knowledge captured in the work environment and find work request(s) that are potentially associated with a given code change. This involves Natural Language Processing (NLP) and programming construct analysis of the outgoing change set, as well as NLP and Orthogonal Defect Classification (ODC) analysis of the open

work requests. Once the highest candidate work items are found, the user has several actions that can be performed on the list of work items.

The system applies several methods to associate code changes to work requests:

• Orthogonal defect classification of defects matching to structured analysis/patterns of the code change set

• Matching the "focus" of the current change set (from main object modified) to a focus within work item(s)

• Consideration of structured information, such as which work requests were previously associated with a change in a given file (e.g., if the last nine changes to a file were assigned to the same work request, then the tenth probably will be, too)

To implement the system for searching problem reports for other defects associated

with a developer's change set:

1. On an ongoing basis, the system intelligently ingests of a corpus of work requests

A. Performs ODC on defects to determine the type of error (e.g.,


Page 02 of 3

initialization, manipulation, etc.)...