Overview of code review

Purpose

Code review is a set of rules that automates the process for a software developer or architect to review code. While the manual code review process can consist of time-consuming and subjective discussions, the automated code review is effective, quick, and consistent. The automated code review supplements the manual code review. It does not replace it.

Benefits

The automated code review tool adds value to the software development process in several ways because it performs the following tasks:

Because the automated process is quick, you can run code reviews often. The code review findings let you catch and correct problems early, when it is easiest and cheapest to make changes.

Supplied code reviews

Several code reviews are supplied. Each code review applies a different set of rules, which are organized in folders. Depending on which stage of the development process you are in and what your review objective is, you can choose the code review to suit your needs. The broadest review is the Complete Code Review that applies rules from all categories, as shown in the following screen capture:

Complete Code Review, expanded to show all categories

Some categories also have a code review associated with them. For example, you could select the J2SE Best Practices Code Review when you want to apply rules for that category only, as shown in the following screen capture below. This lets you run code reviews that focus on a particular aspect of the code.

J2SE Code Review, collapsed

User-defined code reviews

You can create rules from a supplied wizard. The wizard lets you choose from two types of rules: architectural controls and general. These rules allow software architects to extend the capabilities of code review by creating rules to ensure the integrity of their designs.

Severity levels for rules

Each rule has a severity level. For a supplied rule, you can modify the severity level assigned to it. When you create a rule from the wizard, you specify a severity level. The three severity levels are indicated by the following icons:

Although "recommendation" is the lowest severity level, do not discount how important it is to address these findings. They reflect a set of best practices and industry standards that engineering teams should adhere to. Even if these findings are not immediate problems, they could lead to problems in the future.

The following screen capture shows rules in the Comparison folder of the J2SE Best Practices Code Review. Rules in the folder have all three severity levels.

text

Automated fixes for some problems

Some common findings come with a quick fix, which is a supplied automated solution. If a finding in the code review has a quick fix, it is indicated by one of the icons in the following illustration:

text

Summary

In the software development life cycle, code review automates the peer review process on a body of code. The supplied code reviews allow you to run the following types of reviews:

You can also use a supplied wizard to create your own rules that are specific to ensuring the integrity of your application's design structure.

Because automated code reviews are quick to run, you can detect problems and inconsistencies in a code base early. Consequently, you can fix these problems early before they affect your application's maintenance, scalability, and performance.

Now you are ready to begin Exercise 1.1: Importing the required resources.

Terms of use | Feedback
(C) Copyright IBM Corporation 2000, 2005. All Rights Reserved.