Jameleon is a data-driven automated testing tool that is easily extensible via plug-ins. Features of applications are automated in Java and tied together independently in XML, creating self-documenting automated test cases.
The learning curve is small and maintenance is minimal because Jameleon is broken up into different layers that can be learned by people of different skill sets. Jameleon breaks applications up into features and allows testing at any level, simply by passing in different data for the same test. And because Jameleon is based on Java and XML, there is no need to learn a proprietary technology.
Jameleon can also be used be used to generate the test case documentation based on the grouping order of the functional points inside the script. In this tool, when a functional point's script or documentation is updated, the test case documentation also automatically gets the updates.
In order to develop Jameleon in the shortest time possible, Jameleon uses several independent open-source projects:
A functional point is a feature of the application. Some examples of features might be search, login, signup, and edit preferences. Because there is much more to those features than just a simple description, it is recommended to categorize the functional points into three types for maximum reuse,:Please note that separating these functional points is only a recommendation.
|An action point can be thought of as the act of doing. Let's
take search for example. Usually search only has two UI elements to it, a text
box where the search term(s) are entered and a search button that submits the
search term(s). The action point doesn't necessarily care about
the search results. It only cares that the search page was loaded correctly, the
search term was inserted and the search button was clicked.|
In other words, a action point doesn't validate what happened after the action occurred, it only performs the action.
|A validation point verifies the outcome of the action point. Usually an action has several possible outcomes. For example, login might behave differently depending on whether the username and password are correct, incorrect, too long, too short or just non-existent.The action point is the act of logging in and the validation point verifies the outcome when using valid, invalid, too long, too short or just non-existent usernames and passwords.|
|Navigation points traverse to the action point or the validation point so it can be executed. For example, there are websites that have a navigation bar on the top no matter which page is loaded. In this case the navigation point simply clicks on the link in the navigation bar while on any of the pages in order to load the page to perform the action or validation.|
A session is the means of grouping together functional points of a single application. The session keeps state and tracks variables of all the functional points grouped in a single session.
A test case contains one or more Sessions. For example, maybe you want to test the delete user
action point in the admin interface of an application that allows customers to login. A valid
test might be to:
Description Functional Point Type Session Log into the site with a valid customer preferences. Action Point Session 1 Verify the login succeeds. Validation Point Session 1 Navigate to the admin application. Navigation Point Session 2 Log into the admin interface. Action Point Session 2 Validate the login succeeds. Validation Point Session 2 Search for the customer. Action Point Session 2 Validate the customer to be deleted is in the results of the search. Validation Point Session 2 Select to delete the user. Action Point Session 2 Verify the confirmation page is displayed. Validation Point Session 2 Confirm customer deletion. Action Point Session 2 Navigate to customer login page. Navigation Point Session 3 Login as deleted customer. Action Point Session 3 Verify failed login. Validation Point Session 3
In this example, the first trial of logging in with a valid customer would be in one session. Deleting the customer in the admin site would be another session. Verifying the customer can no longer log in would be a third session. Thus the test case would contain three separate sessions. Also note that in this scenario the customer login action point is used twice, but two different validation points are used to verify the outcome.
While the functional points are written in Java, the test case is written in an XML scripting language. Seemingly unrelated functional points can be tied together in one or more sessions with minimal effort with the scripting language, . A tester with a limited programming background can develop the functional points and a tester with no programming knowledge can then group those functional points into a test case.
Testing a functional point is not as simple as just running one test against it. Usually a functional point has rules and boundaries that define failures and successes. With Jameleon, test cases can be data-driven. For each row of data, the test case is executed once and it isn't successful unless each row of data passes. For example, testing a "change email" functional point can't be done with one test. A valid email address adheres to several restrictions. For example, it can't have spaces, quotes, colons, etc in it. The only way to test the "change email" functional point, would be to run the same test case multiple times, each time with a different email address that has a different invalid character in it. Jameleon does this without requiring the script to be changed in any way. In fact, a test case can be run by using a data file or by defining a single set of variables right in the test case itself.
Sometimes the same application exists in multiple Organizations. The application deployed to different
organizations has the same functionality, except the application is customized for each organization
For example, a company owns several sites that each sell different products. The same application or
shopping cart exists on each of those sites, but the customer discounts are different. The only major
differences are simply the look and feel and the products sold on the sites.
Why write a separate test case for each site when the only difference is the data that is validated? In Jameleon, test cases can be scripted independent of an organization. The Organization is defined outside the test case and Jameleon can be told to run the Test Case against any of the organizations the test case is configured to run against.
Most medium to large sized organizations have many environments their applications run in. Usually, an Organization will have a sand box or development environment, a testing environment, possibly a staging environment and obviously, a production environment. Except for a few differences, the applications work the same in each environment. A web application, for example, would probably have different host names where the application can be accessed. The data the application uses might be different. Usually, however, the functionality is the same. Jameleon takes these subtle differences into account and allows a test case to be run in many different environments without changing one line of code.
Jameleon is an object oriented automation testing tool that does not have many of the weaknesses that are inherent in automated testing. The learning curve is small and maintenance is minimal because Jameleon is separated up into different layers that can be learned by people of different skill sets. Separating applications up into features, allows testing at any level, simply by passing in different data for the same test. And because Jameleon is based on Java and XML, there is no need to learn a proprietary technology.
Jameleon is just the beginning of a more advanced tool that will be used to generate the test scripts and test case documentation based on the grouping order of the functional points. In this tool, when a functional point's script or documentation is updated, the test case documentation will also automatically see the changes.
One of the hardest things with automated testing is interpreting the results. Sometimes, testers will spend most of their time looking through cryptic logs to find out why the test failed. In Jameleon, the results are categorized into importance levels. While one resource gathers only the very important results, another resource gathers all results. The results are recorded in a human-readable XML format. The person writing the functional points need not worry about the results being recorded, Jameleon keeps track of everything from execution time at each level to the sum of each of the asserts run, passed, failed and errored at each level.