The architecture we discussed has these main features:
* Operation from a windows base with eValid as the
* Application mode recordings of the Windows app
that run natively.
* Use of connectivity software to view the operation
of your application all of the other platforms, form the
windows platform. If you do this correctly (screen
size and font size) your functional tests in application
mode probably will apply across UNIX platform lines
(but this is Application Mode, not Object Mode).
* Batch operation of eV.Manager ought to let you
run entirely "hands off".
I think this architecture will serve you well and will
greatly simplify the QA task.
Let me know if we need to schedule time to discuss
this approach with others there to get it approved.
If you check on Google for "connective Software Solutions"
you'll find lots of alternatives. Hummingbird is one of them.
Here are the items I mentioned in our call:
* The eValid string substitution facility -- used in both functional
testing AND in server load creation -- is described here: http://www.e-Valid.com/Products/Documen ... rings.html
Here's a diagram of this process: http://www.e-Valid.com/Products/Documen ... nsion.html
* To have one script run repetitively, after adjusting values of parameters
in your script, you can use the powerful eValid Data Synthesis capability
(this capability works in LoadTest runs also): http://www.e-Valid.com/Products/Documen ... .data.html
* There is a built-in capability to pre-specify values for
parameters in an eValid script, described in detail here: http://www.e-Valid.com/Products/Documen ... ables.html