Travel: Problem list
Christian mogensen@cs.stanford.edu
http://www-pcd.stanford.edu/mogens/377/travel.prob.html
Presentation Problems
- The radio buttons on the right of the map are not labeled -
presumably changing the selection will change the map in some way.
Reason: fails to make visible the purpose of the buttons. (Norman)
- The F and C checkboxes presumably relate to the selection among the
radio buttons. (Fahrenheit/Centigrade?) Do they change when the
radio button selection changes? Reason: ambiguous layout -
position does not make interior connection visible. (Norman)
- Problem: Fahrenheit and Centigrade should be radio buttons (mutually
exclusive) rather than check boxes. Reason: source of errors if
both selected simultaneously - should be forced to choose only
one (Woods/Norman).
- The date in the upper right is unlabeled. Reason: ambiguous unmarked inputs are a source
of error (Woods).
- The title is distracting - large black area above window contrasts the
main window. It dominates the window without providing useful
information (original hueristic based on CS147).
The Input Controls
- The date is guilty of not making internal states visible. (Norman)
If we can only select the next two days, then only those two
should be available. We might even mark them "today, tomorrow,
the day after".
- Similarly the time can only be one of four states and does not need
to be a text input field at all.
- The date and time input is not mapped onto a time line or any form of
incremental input even though the underlying data type, time, is a
continuum. (Norman)
- In any case: showing an error box when the input date does not match
the requirements is not on. Making the user agree it's OK is
rubbing salt in the wound. It's a disproportionate error message for
a small error. (Woods)
- The error message also does not say what is wrong, just mentioning
the fact that something is wrong with the weather data. (CS147)
In fact, the error message is misleading since it does not indicate
that the date is out of range. (Woods)
- Beeping on invalid inputs is endurable - not too heavy an error
indicator. Plus it gives feedback to the user indicating that
the program is not ignoring inputs unintentionally. (Norman)
- The error dialog box for the
Unknown map coordinates
has
the same problem as the previous error message: it does not
give any indication as to what is wrong with the input.
This means increased cognitive load on the user who has to
remember what format of input goes where.
- Ditto also on the OK button - a "TRY AGAIN" might be more appropriate.
- Using the keyboard for these inputs while using the mouse for other
selections splits the user's focus, puts extra load on the user to
remember what is being typed and forces a movement from keyboard to
mouse. (Card, Woods, CS147)
- The map input forces you to remember the format of the input -
the format is not visible to the user. (Norman)
- The admirable attempt at feedback fails. Changes to the input boxes
take effect immediately after input is complete (and correct?).
The specification does not mention if this also applies to the
undescribed Temp/Precip/Visibility/Wind buttons. Inconsistent
controls lead to errors as they increase the user's cognitive load
(Woods).
- The input boxes Magnification and Map Center are poor
mappings onto the two dimensional map (Norman). Magnification
is arbitrary and Map Center slightly less so since that at least
is standardized.