Browse Prior Art Database

Automated Prioritization Model for Defect Validation

IP.com Disclosure Number: IPCOM000211179D
Publication Date: 2011-Sep-26
Document File: 3 page(s) / 33K

Publishing Venue

The IP.com Prior Art Database

Abstract

Disclosed is an automated prioritization model for defect validation. The method determines the relative verification/validation test effort for a defect by using key statistics about the defect and the underlying code.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 48% of the total text.

Page 01 of 3

Automated Prioritization Model for Defect Validation

This disclosure presents a programmatic way to determine the relative verification/validation test effort for a defect by using key statistics about the defect and the underlying code. The result of the invention is a prioritized list of defects that indicate to a verification engineer the priority (i.e., care/impact level) associated with defect changes; the first defect in the list requires the highest focus and the last requires the least amount of focus.

Every defect must be verified to insure the original problem was fixed and no new problems/regressions are introduced.

One defect assessment model states different defects have different sizes where size is determined by the effort to verify them. [1] That model discusses distributions of defects, but does not specifically address how to prioritize individual defects based on verification effort calculated with key attributes and automatic examination of the code base.

The disclosed invention provides both a process and measurement tool to perform the sizing of effort described by previous models, where the assessment can be automated, and a final ranking of defects can be used to both prioritize and size verification efforts. "Large sized" defects hold the most potential for regressions, new defects, or failure to fix the initial defect. The results from the invention can be either presented to a human in a report format or processed by an automated system, keyed to understand which test cases to run based on the defect prioritization and modules of code that were changed.

Following are other documented methods to measure defect verification/validation efforts. Each is differentiated from this invention as described below:


• Measuring straight line code paths using cyclomatic complexity, which can estimate the initial tests needed, but does not speak to the effort of re-testing once a defect is found [2]


• Severity/Priority method measures the impacts of the problem, but the effort to verify is not provided by these measures


• An approach that discusses function points in code, including the number of function points total and the number changed.
[3] This does not offer a prioritization scheme for retesting defects based on an estimate on how long it might take to verify a defect fix


• Method using 100's of data points to assess complexity/prioritization and quality, but is difficult to calculate and apply to individual defects. [4] This invention offers a simplified algorithm that can be applied to a discrete set of defects as opposed to a statistical sample of defects. The article also does not specifically address verification of fixed defects but rather finding initial defects.

PDF Creator - PDF4Free v3.0 http://www.pdf4free.com

1


Page 02 of 3

Specific differences/advantages over the other solutions include:


• This invention is not post-verification based, and is practical for measuring individual defects


•...