How do folks coordinate tests running on different machines? Some of our products require tests which, for example, invoke an action on machine A and then perform a verification on machine B.
As a simple "mind experiment", consider testing an Instant Messaging program which communicates between two different PCs. You type "Hi there" on machine A and need to verify that "Hi there" shows on machine B. How would you use QTP to do this? Our testing requirements include physically different machines (e.g. Vista vs XP, different hardware configurations, etc).
I've done searches on this forum and others, as well as Mercury's Knowledge Base (KB), for terms like "synchronize", "multiple machines", and others with no luck. Some close calls:
Mercury's KB Article ID 23695 discusses Synchronization Points, but they seem to operate only on the current application on a single machine.
Article ID 42288 suggests that the BPT product is also needed. Can't multiple machine tests be performed with just QTP?
I haven't learned about the AOM (Automation Object Model) yet, but as a developer that seems like it might hold some promise. Anyone used it for this?
Article ID 44597 suggests writing data to a flat (text) file. For multiple machines that would mean writing to a file shared on a network. Does anyone really do this? This approach seems really crude and clunky to me (but then I'm a software developer with lots of client/server experience). This file-based approach is not event-driven, which means that tests will have built-in delays due to polling.
When we were evaluating vendors we asked Mercury directly about this, since Segue's SilkTest has support for this built-in. Mercury was up-front that they did not have it. So what do QTP folks do? BTW, we are not using BPT, but do have QC.