Concerns about testing a tricky AJAX web page

Applying eValid to AJAX applications that require advanced DOM-based methods.

Concerns about testing a tricky AJAX web page

Postby VickiC » Thu Jan 15, 2015 9:44 am

Morning.

If I use the structural testing methods you have in eValid, is there any guarantee that the element(s) I find at one instant stays in the same place during the same test?

I'm trying to test a tricky AJAX web page.

TY
VickiC
 
Posts: 1
Joined: Wed Jan 14, 2015 3:41 pm

Re: Concerns about testing a tricky AJAX web page

Postby eValid » Mon Jan 19, 2015 3:59 pm

VickiC wrote:Morning.

If I use the structural testing methods you have in eValid, is there any guarantee that the element(s) I find at one instant stays in the same place during the same test?

I'm trying to test a tricky AJAX web page.

TY


Thanks for asking VickiC.

This is a very interesting question because it brings up a point that is often forgotten or missed:

The contents of an AJAX page can vary dynamically at any time and that makes such pages quite difficult to test.

That's just the nature of AJAX pages. Can't do anything about that.

So you are right: there ARE cases when even a structural test that only relies on the layout of the page will fail when something upon which it depends changes.

For example, you may have a drop-down list and you want to click on the items in the list that is named "foo" so you can go to the place that clicking on "foo" takes you. But the page is being updated and for some reason there is a second "foo" entry on the page and now your test takes you -- as you expected -- to the location named "foo" but it probably will be the first one, and probably be the wrong one.

In this case, the only way you could detect that your test's PASS was a false positive ("test passed but it wasn't supposed to") is to do a secondary analysis to confirm that the "foo" it went to was the ONLY "foo" it COULD have gone to. But to do that requires some fancy programming -- probably you need the eValid Programmatic Interface (EPI) for this.

You can see that at some point the complexity of engineering a 100% reliable test for an AJAX page will involving a level of effort that is WAY out of proportion to the value of such a test. Except of course if $1 Million is the cost of not having a false positive.

It is just a matter of economics then.

-- eValid Support
eValid
 
Posts: 2392
Joined: Tue Jan 01, 2008 12:48 pm
Location: USA


Return to AJAX Applications & DOM Processing

Design Downloaded from free phpBB templates | free website templates | Free Web Buttons