Quality Dashboard

Introduction to the quality dashboard

The goal of the Quality Dashboard is to provide a quick overview of the quality of all projects in the Software Factory.

All active work projects and active versions are available. Base projects are excluded.

The dashboard can be found under two menus:

• menu Quality > Quality dashboard - The average quality of all projects together, plus data per project.
• menu Projects > Projects overview > tab Quality dashboard - The quality of only one project and its active versions. To see if this project's quality improves: deselect the Hide inactive project versions prefilter to view previous versions of the project.

note

Is the Quality Dashboard slow? Then unmark the Work project check box for every unused project and set the Status to 'inactive' for all unused versions.

This chapter provides an overview of the quality dashboard's data and possibilities.

Overall quality

This data is only available under menu: Quality > Quality dashboard.

The Overall quality shows the average quality of all active versions of all non-base work projects. Branch-snapshots (for branching/merging) are excluded.

The projects used for calculating the overall quality are listed under Project. If another project is selected, this data will stay the same.

Overall quality

Specification statistics

This data is only available under menu: Quality > Quality dashboard.

The Specification statistics show your work's progress instead of its quality. This data is obtained from menu Specification > Iterations. It provides information on Specification progress, Development progress and Acceptance progress.

Specification statistics

Project

This is a list of all projects being used for the quality calculations. Only the active versions are shown.

To view the inactive versions, deselect the Hide inactive project versions prefilter. This way the quality improvement between versions can be monitored.

note

Note that this can severely slow down the performance, so it's best not to do this in menu Quality > Quality dashboard, but only in menu Projects > Projects overview > tab Quality dashboard

Project

Model quality

Model quality shows whether validations with messages have been solved sufficiently. More unsolved messages lead to a lower number of stars.

Validation

The Validations per severity chart provides an overview of the number of validations, with or without message, per severity type (information, warning or error).

To improve the model quality, check the validations and solve or approve them via menu Quality > Validation. Start with the most severe messages. So, first errors, than warnings and finally information messages.

Validations per severity

Code quality

Code quality shows the quality of your code divided into four categories:

• Development: the average number of control procedures in development per developer. To improve the quality, make sure that developers don't work on too many control procedures at once (less than 5) and that a control procedure is fully completed after finishing.

  A score is assigned immediately at the start of a new project, even if no code has been written yet. The reason is that included base projects are also taken into account.

• Review: the total number of control procedures in review. To improve the quality, make sure to review each other's work as soon as possible and fully complete the control procedure after finishing.

• Reduction: how often a control procedure has been weaved, 25 percent or higher is considered as high quality. Reusing templates leads to a higher quality software, because it reduces the chance of errors.

• Documentation: the amount of comment in the code for program object items, 10 percent or higher is considered as high quality. Stars are awarded for each template separately, here the average amount of stars is shown. Comments contribute to the code's comprehensibility. An overview of the comments in the code can be found in menu Analysis > Templates > tab Cube. Right click in a form and select menu Cube views > Documentation > Ratio code / comment. To improve the quality, add comments where they lack most.

  A score is assigned immediately at the start of a new project, even if no code or comment has been written yet. The reason is that included base projects are also taken into account.

Code quality

The Code analysis chart provides the same information visually:

• Status: review, development or completed
• Reduction: reused code vs unique code
• Documentation: the sum of all comment lines vs the sum of all code lines. Although the ratio might look good here (eg 10 lines of comment for 100 lines of code), separate templates can still lack the right amount of comment. This will be reflected in the number of stars.

Code analysis

Test quality

The Test quality shows the use and success of the available unit and process tests.

• Unit test success: the percentage available unit tests with a successful result. To improve the quality, use the unit tests on a regular basis and solve any problems when they occur.

  Five stars are assigned immediately at the start of a new project, even if no unit tests have been build yet. The reason is that only the unsuccessful tests are taken into account, and no unit tests means no unsuccessful unit tests. The actual amount of unit tests is rated at the Unit test coverage.

• Unit test coverage: the percentage non-generated control procedures (only Type: 'Program object item') that is covered by a unit test. To improve the quality, build tests for more control procedures (Type: 'Program object item').

• Process test success: the percentage available process tests with a successful result. The quality improves if more tests are successfull.

  Test quality

  The same information is displayed in a chart:

  Test quality chart

Role quality

Role quality shows how well roles are being used, based on the average number of menu items per role. The less menu items in a role, the better the quality.

A role should preferably contain only one menu item, more menu items might possibly lead to a security risk.

The Menu items per role chart shows the largest roles in your application (with a maximum of 5 roles). To improve the quality, reduce the number of menu items in the largest roles.

Five stars are assigned immediately at the start of a new project, even if no roles have been defined yet. The reason is that large roles lead to low ratings. So, no roles means no roles that are too large.