Re: First step in changing the labels in WLUS
As you can see, you asks bigger questions than it seems:
torsdag 8. april 2004, 15:56, skrev du :
> Why is this called a test case? what can be tested?
We are giving respons of reading through the labels, and giving
sugestions to test out rewritten labels in a new increment of WLUS.
> is this intented to be a suggestion on which other experts on
> user interfaces can comment? then i would wait for comments.
Unfortunately, when asking for "ekspert opinion", you overjudging, and
introuduce a buraucratic torrent-oriented methodology. I had to
introuduce that perspective, and the problems you get with this
"stopping the process" question, since it was brought to our attention.
The whole point is that ther users are the expert when using a user
interface. The experts should ask the ordinary users what they
experienced when using different parts of the user-interface. Then they
shall listen, not speak :-)
A litle about who are doing the testing now.
3-8 people has spent their whole afternoon from 15:00-23:00 and some
have tested to 24:00 in the night in a period of 15 days. Since they
are not so drilled in giving feedback on user-interface-testing, I have
collected their responses with a litle bit of help from Herman. A
GUI-expert should do just that when testing the usability (its many
more ways, but this is almost bulletsafe). When testing usability it's
the regular user who are the expert, not the persons we ordinary has
that position.
When regular users take the burden on commenting on the user-interface,
they often do that in a less precice way. They can also beleeve that
other persons are experts in designing good user-interfaces when they
are not. When the developer are not aware of this fenomena you get into
this problem:
You overdjudging regular users in a fearly simpel label-issues, as the
ordinary users was asked what they think, and then gave comments with
cange in mind. When the developers ask for other expert opinion, and
second guessing the feedback, the developer breaks the enhusiasm for
giving feedback. If that happens (and it has not), it's realy bad. It's
a lot of science literature on this issues, and I can send you a whole
lot of links on this matters:
http://www.ics.uci.edu/~grudin/Papers/IEEE91/IEEE91.html
Luckily Herman has som experience with GUI-testing at Opera, and how
this is done in practice. I have a major subject in GUI-design with
Jonathan Grudin[1] as a teatcher. I have also worked with this things
at my professional occupation in 3 year before Skolelinux. So we have
people that has conductet this work with me as a test-leader, as I have
done many times before in my profesional occupation.
[1] http://www.ics.uci.edu/~grudin/
The other persons that was joining in was Harald Tingelstad,
Jim Oksvold, and Morten Arnesen. We also got some help from som
yongsters that has greate interest in computers. Also Vidar Bakke has
tested the GUI.
The experts here is Herman and my self. We have just asked the users
what they expect, and collected a lot of feedback on these labeling-
and placeholder-issues along the way. It's not rocket sciens. It's
basic things that could be fixed right away.
> otherwise we could just start changing the texts in the lang/en
> file in cvs.
Give it a shot straight away :-)
Conlusion:
It's a huge difference in second guesings usability and a programable
function from a computer experts. When designing user-interface the
users are the experts :-). When programing business functions the
business-people are the experts, lower down in the application it's
the computer programers. The experts and programers has to know this,
or else this will widens the gap between developers and users as
Jonathan Grudin has written in his classic artikel about the matter[1].
So just change the labels and some placeholders, and se what comments
you get. Use the feedback from the real-life, and change at once. Our
feedback is not from the GUI-experts, but from the real GUI-experts,
the users. If it is for the worse, and a lot of users complains, just
change it back with som CVS-commits.
mvh Knut
Reply to: