[RFC] Color usability time...

[Robert Knight]

Hello Matthew,

Just a word about prioritisation.  Your post discusses quite a variety
of topics.  I think the addition of new colour roles is far and away
the most important item in the list - so perhaps we should concentrate
on that for now.

In particular, the existing colours only handle basic widget  /
subwidget types and a few states.  What is missing is a set of
KDE-wide colors for more abstract things:

For example, colors to indicate:

- Success
- Failure
- Warning
- High Importance
- Low Importance
- New
- Old
- Activity
- Inactivity
- Personal ( something targeted specifically at the user - eg. an IM comment )
- Impersonal ( something not specifically targetted at the user )

I am going to be possibly controversial here and suggest that we don't
rush to provide a configuration tool with KDE to change all of these
colours individually,   Instead with the focus on trying to provide a
selection of attractive and tested colour schemes, including several
to cater for common problems which people may have with colours.
People who understand accessibility better than I do please comment if
this is not-doable.

One other thing is that there are times when it might an application
might require a use which is not covered in the list.  Perhaps we
could provide a method for an application to ask whether a particular
colour combination is acceptable ( ie.
KSomeClass::areColorsAcceptable(foregroundColor,backgroundColor) ) and
if not, the application can fall back to something safer.

Regards,
Robert.

On 04/06/07, Thomas Zander <zander at kde.org> wrote:
> On Monday 04 June 2007 20:42:31 Matthew Woehlke wrote:
> > == HSL and HSY
> > In order to most closely conform to human perception of color, we
> > really should be doing operations in HSY.
>
> The reason Matthew says the above is because he is of the opinion that the
> following screenshot shows one bar for each transition that is absolutely
> of lesser quality (and unusable for the usability principles) then each
> other. (which one do you like best)
> http://lists.kde.org/?l=kde-core-devel&m=118013796925437&q=p4
>
> Naturally we are not doing a science class and we surely are not using
> perfect equipment[1] so we should really question the authority with
> which he states these supposted facts.
>
> 1) a CRT monitory typically has a 20% color inaccuracy (variance). Laptop
> screens are worse since they can have different backlight levels to top
> it off.
> --
> Thomas Zander
>
>