Application evaluation.

Corey L. Carlson (corey@itlabs.umn.edu)
Wed, 25 Feb 1998 19:07:31 -0600 (CST)

From: "Corey L. Carlson" <corey@itlabs.umn.edu>
Message-Id: <199802260107.TAA26601@trout.itlabs.umn.edu>
Subject: Application evaluation.
To: beec0018@tc.umn.edu, mcco0062@gold.tc.umn.edu, nord0157@maroon.tc.umn.edu,
Date: Wed, 25 Feb 1998 19:07:31 -0600 (CST)

Design Evaluation
Corey Carlson corey@itlabs.umn.edu

Overall I was very impressed with the system. The use of realistic
miniatures made user feedback very clear and helpful. The only real
problem that was not able to be resolved was the function of the save
versus the apply button. If this cannot be avoided then perhaps a
tooltip or something that explains their difference would be useful.

Task 1:
Change the title bar color to red.

This task began with a click on a button labeled unambiguously. I
don't remember the label, but it made the choice among the four
easy. There is a blank square that has the word "tool" in it that is
quite prominent. Given its size it should afford something, but I
didn't see the relevance of the word tool and could not make it do
anything. I then assumed it was cosmetic and moved on. The next window
brought us a MS windows like miniature for adjusting options related to
the windows. Given some MS windows experience the miniature afforded clicking
upon the representation of the item that required change. At first I
didn't realize that text color was modifiable so I had selected the
text instead of the window border for color change. This was reflected
by changes in the selection in list boxes at the bottom of the screen
so it was possible to detect my error. The next step was to click on
the miniature of the title bar, but not in the title label. This
presented appropriate options in the list-boxes below. This feedback
allowed me to recover from a mistake and know when I had selected the
right item. The next step involved a pretty standard color picker. The
reliance on standards made this a trivial step. The miniature provided
good feedback by changing the title color to match the selection from
the color picker. Finally came the last question. Do I save or apply?
This created a bit of a gulf in execution as we were hesitant to
choose either when not having a good indication of the results. In the
end we chose apply and felt we were done.

Task 2:
Change the screen-saver to soccer balls and set it for 10 minutes.

The first step was to click on the screen-saver button. In the
resulting screen a list-box of possible screen-savers appeared on the
bottom and a miniature preview of the screen-saver appeared to the
right. A labeling problem ensued when we could not find one labeled
soccer. An inspection of most of the items in the list-box did not
provide a good clue as to which one had to deal with soccer. The
correct choice was labeled with football (quite correctly if outside
North America). When selected it yielded a miniature with bouncing
squares instead of soccer balls. It took a guess that the miniature
that provided feedback on the selection may be in error and a click on
the preview button which activated the screen-saver full size. This was not an
intuitive jump to finding soccer balls and many may have been left in
a gulf of execution wondering what to try next. The next step was to
set the time for 10 minutes. This was a bit confusing for me because
the text widget was much larger than the entries for the list boxes. I
kind of expected all of the text to be the same size. This may be a
consistency problem. After clicking in the text box it was clear what
its purpose was so entering 10 was the natural solution. Finally there
is the same problem with apply and save at the end of this task.

Task 3:
Remove the third button down from the button bar at the right of the
screen.

The first button was once again well labeled and brought up the
section for dealing with buttons on screen when clicked. There was
once again a good miniature model of the screen displayed. This
realistic miniature allowed us to click on the button to remove
(number 3). At this point number 3 became selected below in the
options as a feedback device as well as the third button on the
miniature changed to reflect selected status. Then we clicked on
delete. We were asked to confirm the delete (verbally since it was yet
unimplemented) which gave us one last chance to change our mind. This
may have been in place to guard a non reversible step. Saying yes to
the dialog and clicking apply (since we applied in the other 2 tasks
and were not sure of the function save provided) finished the task. It
is notable that if one were to use this often it would become second
nature to click OK and the warning would be wasted.