Presentation Notes

Amir Kolsky - What do we mean by device independence?

W3C Workshop on Web Device Independent Authoring
Wednesday 4 October 2000

Notes taken by Gregory Rosmaita

Background for Breakout
Amir: What do we mean by Device Independence

channels and modalities (high capacity/low capacity)
features and capabilities within a specific modality --
voice capabilities, keyboard, graphical capabilities;

who's responsible for defining DI within sub-sets? an
author would like DI in one modality? is that what we
mean by DI? one channel with lots of features? DI
across multiple channels? who is responsible?
pragmatic POV, much easier to provide DI in one
channel, going across many channels difficult

how do we achieve DI? do we provision data using
markup language or authoring tools? burden on
author's logic? is the channel responsible for this
(push generation -- how do you know how to format
it? only when delivered; should the device have a set
of profiles? how do we formalize this?

DI -- content? UI? both? how do we capture both?
can markup content to display nicely, but UI a
different issue; should they be under the same
scope? if mix them both, how do we do it? is content
in charge is application in charge?


updated slides and graphs

4 possibilities
proximity to provider or device
publishing cycle
programming components
channels and capabilities

Q: the author or the browser? both are responsible,
but which comes first? writing the accessibility into
the browser or the author into the pages? it has to be
both; a reference model is definitely something that
we need to do
A: reference models are very important to telephony;
they are interfaces which are well known; if write to a
specific reference points (where 2 pieces of software
interact) -- can talk to each other if have same
reference points; abstract reference points are what is
Q: we used to say in accessibility business, "until
things work in the ideal world, let the user or user's
device have the last say; if author insists on complete
control, what is to be done? as much as possible can
be decided by UA preferences; content provider in
multi-modality provider should be most concerned
about loss of content
A: just basis for discussion; presenting this to stir up
Q: in the TV environment, there is a tension btw
application /programmatic approach as opposed to
declarative style of design; accessibility issues are
tractable in declarative format; in application format,
have to educate each individual program, because
once the program is written, you can't change it; can't
be retrofitted at browser level