command access via (multiple / no) devices

At 12:08 PM 12/16/98 -0500, Charles McCathieNevile wrote:

>I was basing my statements on the revsed wording which I proposed - device
>independent control mechanism, rather than redundant means, since I don't
>see that there is a need for much redundancy, which simply adds bloat to
>the progams.

That is the fundamental and fatal flaw in your reasoning: you are proposing
to substitute apples for oranges.  A programming interface is not a user
interface.
Device-independent user access to control capabilities, including
meta-control capabilities such as profiling the user interface, is an
oxymoron.  You can't give the user access to a verb without an interface
device.

A given software product will support some user interface profiles natively
and be adaptable to support other user interface profiles.  Only in the
case where "adaptable to support" capability is claimed does the inteface
to the rest of the UA matter.  But in the case of UI profiles supported "by
adaptability" the adaptive add-on forms part of the UA and we have
certainly violated the "please, no assembly required" request.  The rest of
the UA, supplied by the add-on, accesses devices to communicate with the
user.  The requirement to support diverse devices for each verb is the
fundamental requirement, and it can only be verified in terms of the actual
devices accessed by the UA including the add-on components.

Al

Received on Wednesday, 16 December 1998 13:19:23 UTC