Linked by Thom Holwerda on Tue 27th Jan 2009 13:46 UTC
Editorial With Windows 7 having made its grand debut, and with KDE4's vision making leaps and bounds forward with every release, we have two major software projects that have decided to implement some fairly drastic interface changes. Such changes are bound to receive some harsh criticisms - but the funny thing is, these criticisms usually come from people you least expect it from.
Permalink for comment 345831
To read all comments associated with this story, please click here.
RE: Aack!!
by Moochman on Tue 27th Jan 2009 21:41 UTC in reply to "Aack!!"
Member since:

I hear a lot of people discussing how bad SUSE's "start" panel change they made to Gnome is... and how bad the YaST and Compiz setting manager panels are... BUT supposedly, these were designed based on countless trials with ACTUAL USERS.... so the UI is "good" because of that, right?

This is the interesting thing about usability testing... it doesn't always lead to the optimal solution. Usually this is because in order to do said testing, the implementation must already be in place. If the said implementation isn't especially optimal to begin with, then no amount of usability testing will fix it.

I don't know the details of the SUSE testing, but my guess is that they pitted the old start panel against the new one and tested to see how fast users performed various actions and what kinds of feedback they got. They thereby found that the new menu outperformed the old, and maybe made a few tweaks based on feedback.

Whether the new menu is truly optimal, however, is still unproven--and is in fact impossible to prove via quantitative methods. The best measure, though, is probably the response of the actual users once it's released into the wild.

Of course then if it turns out no one likes it, it's too late....

Which is, suffice it to say, the reason why design *prior to implementation* is so important.

Disclaimer: I'm not saying SUSE didn't design prior to implementation; I have no idea how their process worked. I just felt like going off on a tangent about the pitfalls of relying purely on usability engineering.

Edited 2009-01-27 21:44 UTC

Reply Parent Score: 2