Browse Prior Art Database

AC - System and method for suggesting code review based on learning model

IP.com Disclosure Number: IPCOM000234983D
Publication Date: 2014-Feb-21
Document File: 1 page(s) / 40K

Publishing Venue

The IP.com Prior Art Database

Abstract

Disclosed is are a system and method to train a learning model with the development history of one developer, examining the developer’s code change sets and defects opened up against that developer. Thus, the system can respond with timely and relevant suggestions to active code development in order to prevent mistakes.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 57% of the total text.

Page 01 of 1

AC - System and method for suggesting code review based on learning model

Integrated Development Environments (IDEs) for software development are increasingly complex tools. IDEs are able to offer advice to developers in various situations; however, the IDE becomes bloated as, by default, all advice/assistance tools are turned on.

A function is needed to enable assistance only within the areas needed by a specific developer. This would improve the efficiency of resources and the helpfulness of the IDE suggestion system.

The novel contribution is a system and method to train a learning model with the development history of one developer, examining the developer's code change sets and defects opened up against that developer. The system then determines the kinds of mistakes that the developer is most likely to make or have most recently made. The system consults a best practices model to find advice for that mistake and present it to the user.

For example, if a user produces many NullPointerExceptions (NPEs), then the system provides warnings/advice when the developer creates a new method to handle if any input parameter is null.

To implement the system and method for suggesting code review based on learning model in a preferred embodiment:


1. Ingest a corpus of work items/defects

2. Perform Orthogonal Defect Classification (ODC) to determine the attributes/classifications of these defects (e.g., initialization errors, control-logic flaws, etc.)

3. Determine the f...