RE: Sv: Question from Last Week's WCAG Teleconference

As I have said before, customized user interfaces based on user-selected
preferences are not pie in the sky.  They can be built with today's
technologies.  It's not even hard.  

*	put your content in a database (or XML)
*	build several templates, one for each interface you want to expose
*	use a form to gather user preferences
*	use client or server script to save those preferences to a cookie
(or database record indexed by a cookie)
*	use server script to select the appropriate template based on the
saved preferences
*	use server script (or XSLT if you like) to assemble a page combining
the content from the database with presentation from the selected template. 
*	put a link to the preference-gathering form somewhere visible, so
users can change the preferences when they want to.

The last bullet addresses Claus' concern.  I agree that users should be able
to control what they see. I even drafted a guideline about it <grin>.  I
just don't agree that they should be required to do so constantly.  I've
seen a great deal of usability data showing that most users find constant
decision making to be very difficult, consider "extra" information to be
noise, and are confused by complexity.  I suspect that this would be even
more true for cognitively disabled users.
 
I could build a site like this in a weekend with several brands of
commercially available software.  This is how customized portals, and many
other existing sites, work. 
 
The interesting/difficult part is deciding what preferences to let a user
set.  The customizable portals ask if you want news or sports or stock
reports.  This is not useful for our purposes.  I'm guessing that a lot of
Edapta's value-add is in determining the set of preferences (Kynn, feel free
to correct me).  Voice vs. GUI is probably one, since it is very difficult
(perhaps impossible) to design a single interface that is both a good voice
interface and a good GUI.  Screen size is probably another.  Font size might
be another, or color depth, or reading level, etc. etc..  It might be a
useful task for WAI to develop such a list.
 
It's true that these things are not currently exposed by default in an HTTP
transaction, even when the operating system does collect them.  That does
not mean that a particular site developer can't ASK THE USER these questions
once, save the answers, and then tailor the presentation for future visits
based on that data. 
 
If this data is later available as part of CC/PP or some other standard,
that would be great.  It would make things easier, and promote consistency
across sites.  But it's not required to achieve the effect.

 -----Original Message-----
From: love26@gorge.net [mailto:love26@gorge.net]
Sent: Thursday, October 26, 2000 6:14 AM
To: w3c-wai-gl@w3.org
Subject: Re: Sv: Question from Last Week's WCAG Teleconference



At 02:25 AM 10/26/00 -0700, Claus Thøgersen wrote:


I am worried about the idea that specialized user experiences will solve
many problems because I have never seen a system that is capable of behaving
in accordance with my very unstable and often unprodictable preferences in
other ways than by obeying my commands.


Hence the term "pie in the sky". What's "gonna be" (particularly "just
around the corner" or "real soon now" or "within six months") has a strong
correlation with the term "vaporware". I have no doubt that there will be
systems that *seem* to answer all these issues via CC/PP, etc. but their
deployment is likely to take longer and not be as effective as our dreams of
them - at least that's been my experience with everything ever proposed in
any field whatsoever.

So - in our 50,000 foot view from above it would appear that despite there
being a billion or so pages online, HTML is the same kind of dinosaur as
COBOL (which still probably employs more programmers than C, C++, and Java
combined) - NOT! We're a long way from going through the XHTML transition
and pretending that XML (the problems of a not-yet "new thing" seem easily
surmountable until you try to climb on board a moving train) will make our
current efforts vain if we can just wait a minute (or the now-proverbial six
months).


The upshot is that we are well on our way to a nicely generalized,
reasonably abstract guideline set called 2.0 and what we learn from
hammering at GL 3.1, etc. in the previous incarnation should be heeded. "Do
not use color *only*" isn't parallel to "provide equivlalentcies" and that's
the kind of stuff we must get clear in our minds. The overarching "first, do
no harm" and "erect no barrier" deserve (and are getting) extensive
explanations/examples and we should be able to continue refining these
statements with clarifications.

Claus' "unpredictable preferences" are truer for the device than for the
person involved - and we must not forget that there's a person involved.
This is pretty much a DUH! kind of thing but we all tend to expect inertia
not to be a part of momentum, and that just ain't the case. I've worked over
20 years on a project that is just now starting to fructify. Our stuff won't
take that long, but...

There is a doughnut around that hole and we must keep our eye on it.

--
Love.
                ACCESSIBILITY IS RIGHT - NOT PRIVILEGE