Your e-Business Quality Partner eValid™ -- Automated Web Quality Solution
Browser-Based, Client-Side, Functional Testing & Validation,
Load & Performance Tuning, Page Timing, Website Analysis,
and Rich Internet Application Monitoring.

eValid -- Microsoft Maps Live AJAX Testing Demonstration
eValid Home

Summary
This page presents eValid's solution to testing and validating Microsoft's search and mapping solution Maps Live.

Application Description
This web application is implemented with modern AJAX methods and has some very powerful features:

Test Suite Structure Patterns
The functionality of this website and the way it is structured suggest that a complete regression test -- which may involved a large number of actual test scripts -- would probably involve primarily the following classes of functional tests that all can be run independently of each other:

  1. The ability to search for a particular location and view and validate data about it.
    Script Example #1

  2. The ability to search for a location and show an appropriate map that shows where it is.
    Script Example #2

  3. The ability to manipulate the map display and see and validate different, updated information, based on the newly generated map display.

Demonstration Script Observations

  1. The application "has memory" in the sense that if you make a search and navigate on the map, the application remembers where you last were. We nave noted that this feature interfers with test reproducibility.

  2. To assure fixed initial conditions for our tests, we start them at www.virtualearth.com which navigates to the actual Live Maps search page but does so with with a known (empty) initial state for the application.

  3. An additional requirement of the demonstration scripts will be that they be entirely desktop safe (so that they can be operated in parallel, e.g. in multi-playback monitoring mode or in server loading mode).

  4. If a script cannot be made desktop safe we note that in our documentation for that script.

  5. This application is quite large -- ~2.0 MBytes total download, most of it on the initial page load -- so the behavior of these scripts on a "slower" machine may differ from what we have observed on relatively "fast" machines.