Interaction Design and QA

2 Aug 2006 - 7:14am
8 years ago
2 replies
497 reads
McCarthy, Ann Marie
2006

I'm working as an interaction designer in an agile development
organization. We're making significant changes to the u.i. across a
broad suite of products. Part of the process is to have QA run the code
through automated tests. Some of the design is difficult to validate in
testing--the styling of a button or the behavior of tabs, for example.
Does anyone have experience in working successfully with QA in this
environment?

Comments

4 Aug 2006 - 2:48pm
cfmdesigns
2004

>From: "McCarthy, Ann Marie" <amccarthy at Kronos.com>
>
>I'm working as an interaction designer in an agile development
>organization. We're making significant changes to the u.i. across a
>broad suite of products. Part of the process is to have QA run the code
>through automated tests. Some of the design is difficult to validate in
>testing--the styling of a button or the behavior of tabs, for example.
>Does anyone have experience in working successfully with QA in this
>environment?

I've done the QA side on semi-agile projects in the past.

The biggest problem you'll run into there is that in such environments, the UI design isn't typically set in stone by the time the developers put it together and hand it off to QA. It's hard enough for QA to "black box" test look and feel and usability of an environment going through agile-type changes on a regular basis. Asking them to automate that... no, no, no.

Automated testing tends to only work well in a stable environment. Every change that comes in causes a cascade of "something's changed, look at it, evaluate it, post a bug or revise the tests". If your UI is going to be changing with every build, the automated tests never get to land on something stable enough to create a baseline from. I think you're just asking for headaches, trying to do that sort of testing in that sort of environment.

-- Jim Drew
Seattle, WA

7 Aug 2006 - 4:56am
Adrian Howard
2005

On 2 Aug 2006, at 13:14, McCarthy, Ann Marie wrote:
[snip]
> I'm working as an interaction designer in an agile development
> organization. We're making significant changes to the u.i. across a
> broad suite of products.

If you've not done so already - consider joining the agile-usability
mailing list <http://groups.yahoo.com/group/agile-usability/>. You'll
find folk in similar situations and, hopefully, some useful advice :-)

> Part of the process is to have QA run the code
> through automated tests. Some of the design is difficult to
> validate in
> testing--the styling of a button or the behavior of tabs, for example.
> Does anyone have experience in working successfully with QA in this
> environment?

What are the reasons for it being hard to test? This might be a sign
that there are some of the UI abstractions aren't being represented
explicitly in the code.

Any more info on what kind of application you're developing? For
example if you're developing a web application your testing folk
might want to look at tools like Selenium <http://
selenium.thoughtworks.com/> and Waitr <http://wtr.rubyforge.org/>
that can help test the UI more directly.

(Also, having a separate QA group/phase is raising some flags for me.
In the best agile environments I've worked with the QA people are
integrated with the development team. Some more info on how your
process is set up might help.)

Cheers,

Adrian

Syndicate content Get the feed