Patrick H. Lauke wrote:
Geoff Deering wrote:
You find these types of web environments mostly on intranets. For a
lot of people in large organisations, these are primary interfaces
they have to work with. To neglect to address this issue correctly
could easily impact the integrity of data because the interface is
not communicating *state*, because if the designer is unaware of
this, and overrides this visual communication of state that the user
agent is conveying, with their own arbitrary design implementation,
it would be miscommunicating the state of the data. In this case, it
would be a major design blunder.
But is the solution to make a sweeping "don't style inputs"
recommendation, or to actually educate the designers not to just make
arbitrary decision, but decisions firmly based on usability (including
expected behaviour/presentation of state)? And, assuming that a design
has been implemented which does not exactly match OS conventions but
is nonetheless clear, understandable and usable, is it not just as
valid, as long as the interface design within the pages themselves is
consistently applied?
No, because what you are saying here is a whole heap of criteria that
does not address the priority issue, and that is not doing anything by
design that will override the user agent visually representing the state
of the data (controls) to the user. I'm talking explicitly about
misrepresentation of state.
--------------
Geoff Deering
******************************************************
The discussion list for http://webstandardsgroup.org/
See http://webstandardsgroup.org/mail/guidelines.cfm
for some hints on posting to the list & getting help
******************************************************
