Re: Part II: Issues raised during Mac IE evaluation of UAAG 1.0

David Poehlman wrote:

> a color deficiency is also one which does not allow color to be
> deffinitive in discerning where focus is.  It was implicit in our
> discussion even though it was not necessarily discussed or implied.  I
> however do not take issue with the approach and there are certainly
> other cues that are available.  I only raise it as a point for practical
> consideration when determining whether or not to remove this aspect from
> our document.  I recall also that it it is a part of the wcag.  I am
> still uncertain that we are covered adaquately if we remove this.


In the case of authors, if the author only encodes
semantic information using color, then the user agent
has no way (in a well-defined manner) to communicate
the semantics through other means. If the author
uses red to highlight something important and uses no
other mechanism (e.g., markup or text) to convey that
information, then the user agent can't help the user
who relies on output other than color.

The case of user agents is different: the user agent
maintains state information about focus, selection, and
other UA-controlled parameters internally. In Guideline
six we already require that the UA make available that
information through an API. The UA is also required to
make the information available through its native user
interface, which in the case of colors, implies a graphical
user interface. In the case of UAs, the redundancy is already
there, unlike the case of authored content, which is why
WCAG requires it.

In short:

  * We require the UA to maintain state information about
    focus (and selection, when implemented).

  * We require the UA to make available the information
    about focus location through an API. This will most likely
    be done by handing off a pointer, but could be done with xpath
    or some other mechanism for indicating a range within a tree.

  * We require the UA to allow the user to control the highlight
    mechanism for the focus. If the highlight mechanism for a
    graphical user interface is color, we need to ensure that the
    user can choose preferred colors.

  * I don't think we need to require the UA to use a non-color
    highlight mechanism by default, and certainly not as a P1
    requirement. As I said, it's probably best to inherit the
    operating environment's standard highlight mechanism, which
    may be color-based. Diverging from that may cause more confusion
    because it's non-standard.

  _ Ian


-- 
Ian Jacobs (ij@w3.org)   http://www.w3.org/People/Jacobs
Tel:                     +1 718 260-9447

Received on Tuesday, 2 April 2002 13:50:05 UTC