TECHNOLOGIES FOR CODE FAILURE PRONENESS ESTIMATION

• US 20110161932A1
• Filed: 03/07/2011
• Published: 06/30/2011
• Est. Priority Date: 04/26/2007
• Status: Active Grant

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Corporation
• Inventors: Bhat, Thirumalesh, Nagappan, Nachiappan
• Granted Patent: US 8,332,822 B2
• Time in Patent Office: Days
• US Class Current: 717/122
• CPC Class Codes: G06F 11/3616 using software metrics

Abstract

The present examples provide technologies for estimating code failure proneness probabilities for a code set and/or the files that make up the set. The code set being evaluated is typically comprised of binary and/or source files that embody the software for which the estimates are desired. The estimates are typically based on a set of selected code metrics, the code metrics typically selected based on corresponding failures of a previous version of the software. A historically variant metric feedback factor may also be calculated and code metric values classified relative to a baseline code set embodying the previous version of the software.

Claims

1. A method comprising:

calculating, by a computer, churn metrics that indicate a degree of change between a first version of code and a second version of the code;

calculating, by the computer for the second version, a historical variant metric feedback factor based on first metrics determined from the first version and on the second metrics determined from the second version and on the churn metrics; and

performing, by the computer and based on the second metrics or on the churn metrics, a logistical regression resulting in a code failure proneness probability for the second version.

2. The method of claim 1 wherein the churn metrics include M1 that represents new contiguous code in the second version that is not in the first version and contiguous code in the second version that has changed from that in the first version, relative to code in the first version.

3. The method of claim 2 wherein the churn metrics include M2 that represents contiguous code in the first version that is not in the second version, relative to code in the first version.

4. The method of claim 3 wherein the churn metrics include M3 that represents new contiguous code in the second version that is not in the first version and contiguous code in the second version that has changed from that in the first version, relative to code in the first version, relative to contiguous code in the first version that is not in the second version.

5. The method of claim 4 wherein the code failure proneness probability for the second version is based on the performing the logistical regression using M1, M2, and M3.

6. The method of claim 5 further comprising determining for each of the first metrics a corresponding statistical upper bound.

7. The method of claim 6 wherein the historical variant metric feedback factor is a sum divided by a total number of the second metrics, the sum being a total number of each of the second metrics that is greater than its corresponding statistical upper bound.

8. At least one computer storage medium storing computer-executable instructions that, when executed by a computer, cause the computer to perform a method comprising:

calculating churn metrics that indicate a degree of change between a first version of code and a second version of the code;

calculating, for the second version, a historical variant metric feedback factor based on first metrics determined from the first version and on the second metrics determined from the second version and on the churn metrics; and

performing, based on the second metrics or on the churn metrics, a logistical regression resulting in a code failure proneness probability for the second version.

9. The at least one computer storage medium of claim 8 wherein the churn metrics include M1 that represents new contiguous code in the second version that is not in the first version and contiguous code in the second version that has changed from that in the first version, relative to code in the first version.

10. The at least one computer storage medium of claim 9 wherein the churn metrics include M2 that represents contiguous code in the first version that is not in the second version, relative to code in the first version.

11. The at least one computer storage medium of claim 10 wherein the churn metrics include M3 that represents new contiguous code in the second version that is not in the first version and contiguous code in the second version that has changed from that in the first version, relative to code in the first version, relative to contiguous code in the first version that is not in the second version.

12. The at least one computer storage medium of claim 11 wherein the code failure proneness probability for the second version is based on the performing the logistical regression using M1, M2, and M3.

13. The at least one computer storage medium of claim 12 further comprising determining for each of the first metrics a corresponding statistical upper bound.

14. The at least one computer storage medium of claim 13 wherein the historical variant metric feedback factor is a sum divided by a total number of the second metrics, the sum being a total number of each of the second metrics that is greater than its corresponding statistical upper bound.

15. A system comprising:

a computer;

a code churn calculator implemented at least in part by the computer and configured for calculating churn metrics that indicate a degree of change between a first version of code and a second version of the code;

a historical variant metric feedback calculator implemented at least in part by the computer and configured for calculating, for the second version, a historical variant metric feedback factor based on first metrics determined from the first version and on the second metrics determined from the second version and on the churn metrics; and

a statistical modeler implemented at least in part by the computer and configured for performing, based on the second metrics or on the churn metrics, a logistical regression resulting in a code failure proneness probability for the second version.

16. The system of claim 15 wherein the churn metrics include M1 that represents new contiguous code in the second version that is not in the first version and contiguous code in the second version that has changed from that in the first version, relative to code in the first version.

17. The system of claim 16 wherein the churn metrics include M2 that represents contiguous code in the first version that is not in the second version, relative to code in the first version.

18. The system of claim 17 wherein the churn metrics include M3 that represents new contiguous code in the second version that is not in the first version and contiguous code in the second version that has changed from that in the first version, relative to code in the first version, relative to contiguous code in the first version that is not in the second version.

19. The system of claim 18 wherein the code failure proneness probability for the second version is based on the performing the logistical regression using M1, M2, and M3.

20. The system of claim 19 further comprising determining for each of the first metrics a corresponding statistical upper bound, wherein the historical variant metric feedback factor is a sum divided by a total number of the second metrics, the sum being a total number of each of the second metrics that is greater than a corresponding statistical upper bound.