[Accessibleweb] Minutes of April 19 AccessibleWeb@U - Ginny Redish

Rick Ells rells at u.washington.edu
Thu Apr 22 16:56:03 PDT 2004




Making Web Sites Accessible AND Usable
I. Speaker: Janice (Ginny) Redish ( http://www.redish.net/ )
A. Reported on a study she did with Mary Theofanos
1. Bridging the gap: between accessibility and usability by
Mary Theofanos and Ginny Redish, ACM Interactions,
Volume 10 , Issue 6 November + December 2003
http://portal.acm.org/citation.cfm?id=947227
2. Mary Theofanos is with the National Institute of
Standards and Technology
a. Active in the Industry Usability Reporting Project
http://zing.ncsl.nist.gov/iusr/
II. Two related studies
A. Two different types of disabled users
1. Users who listen to the screen
2. Users who magnify the screen
B. Topics to be explored
1. Navigation
2. Content
3. Forms
4. Other issues
C. Want to go beyond 508 to make truly usable and accessible
sites
1. With users who magnify, found the needs were very
diverse, hard to come up with guidelines
D. How do we achieve "experience equity" and universal access
for everyone?
III. Why do the study?
A. Substantial population worldwide
1. 750 million people have a disability
2. 3 of every 10 families are touched by disability
3. 180 million people are blind or visually impaired
B. United States
1. 21% of Americans currently have disability
2. More than 4 in 10 are online
3. Web is tremendously liberating for many handicapped
4. Have a large discretionary income, at least on paper
IV. The Study
A. Users who listen
1. 16 users tested
2. Work with JAWS ( http://www.freedomscientific.com/ )
or Window-Eyes (
http://www.accessingenuity.com/Product%20Pages/Windoweye
s.htm )
3. 2 hours each, typical usability testing scenarios,
looking at US federal Web sites
4. Took about twice as long as usual usability tests, which
seems typical
B. Users who magnify the screen
1. 10 users tested
2. Working with ZoomText
http://www.synapseadaptive.com/aisquared/Products/zxm
ain.htm
3. 2 hours, same methodology as with listeners
C. Funded in part by US Department of Health and Human Services
1. Basic question to explore: "Is http://hhs.gov/ site
accessible"
2. Real question is "If it is accessible by 508, is it
actually usable?"
a. Just because it has a good score on Bobby, does not
mean it is usable for the people who need it.
V. Goals
A. Understand the relationship between accessibility and
usability
B. Understand how users work with web sites, both users who
listen and users who magnify
C. Develop research-based guidelines for evaluating
accessibility and usability
VI. Findings
A. Users who listen
1. Scan with their ears (they are as impatient as everybody
else)
a. Impatient - just like other users
b. Want to find what they need quickly, but can be
impeded by page design or style conventions
1. Example - list of FAQs where every questions
line begins with "Updated"
2. Example - Alt texts all began with "decorative
bullet image"
2. Skipping the Navigation
a. All wanted to skip the navigation
1. Discovered that they often did not know how to
do that with their software
2. Two jumped to the bottom and read from bottom
to top
3. All sites had a skip navigation link
a. Most did not know about the link
i. "skip navigation" is jargon
ii. "skip to content" JAWS mispronounces
iii. "skip to main content" seems best
3. Software is keyboard oriented
a. Software is keyboard combination driven
b. Software is modal, shift between modes to enter
text, move around, navigate
c. Significant cognitive burden of a product that is
command driven.
1. Must memorize many commands
2. Recognition of command choices is not an
option as it would be in a graphical interface
4. Listening Only to Links
a. Everybody knew how to listen to links
b. JAWs can bring up a window with just the links
c. Example: Looking for "diabetes" when the word
"diabetes" is in a sublist of "Diseases and
Conditions"
1. One problem is many blind are poor spellers
because they have little practice,
a. Screen readers also pronounce words even
if they are incorrectly spelled
b. Can set JAWS to spell the letters out as
you enter them
2. Find in Internet Explorer does not cycle, you
just search from where you are down or up, but
you have no feedback where you actually are
a. Other browsers have better search utility
3. One person canceled the search, rather than
choosing direction, but could not tell what
was happening
d. Can be very quick way to access information, if
number of links is not overwhelming
e. Problems with scanning links quickly
1. Questions are great, but first words in
question do not give clue of what question is
a. Scanning a list of FAQS, yoy may hear
"what, how, where, how, how, how, how,
etc."
b. Real problem
c. Could select better first words
"Volunteer, etc."
5. Jumping from heading to heading
a. Everyone scans the short, noticeable things
(headings)
b. Only works if there are headings, and if the
headings are coded properly (as h1s, h2s, etc.)
1. Saw many pages where headings were not coded
properly (large, bolded text is not the same
as text logically marked with the appropriate
header tags)
6. Occasionally use Ctrl-F - Find
a. Hard to use because you are not sure where you are
on the page
7. Rarely use the Virtual Viewer in JAWS
a. Brings up window that tells you about the page
B. Content for users who listen
1. Do not understand words when the software mispronounces
a. words with more than one pronunciation - content
b. Web words - homepage
c. Unususal words - preparedness
d. Made up words - MedlinePlus, LiveHelp
e. Acronyms - FY (fiscal year) pronounced "fi"
2. Find link text like "Click here" useless
a. Navigate by going down through links
b. Example: Questions are not links, but answer linked
to word "Answer" - all a person scanning the links
hears "answer, answer, answer..."
3. Get confused if the ALT tag and the words on the page
differ
a. What it said on page was "print answer" but ALT tag
on printer graphic said "Printer friendly version".
When wanted to find that location searched for
"Printer" which was not found (not in text)
C. Content for users who magnify
1. With magnification and color change, users may mistake
left navigation for main content
a. People would wipe out any color because of glare or
contrast
b. Color may be only thing that distinguishes
navigation from main page content
2. If content done as graphics, it may not magnify
3. If page not coded well, it may not magnify
a. Fixed font size
4. Users miss things on page, even though they may be right
next to each other, because of magnification
a. Explanatory text separate from entry field
b. Page divided into multiple columns, hard to tell
how far down the page you are
D. Navigation for users who magnify
1. Develop strategies for getting an overview, or not
2. May use scroll bar, but have problems orienting
themselves
3. Very mouse oriented
4. Many magnifiers were working at 5X
a. How do you have any sense of what content you are
dealing with
5. Of our ten low-vision users
a. 3 used same magnification throughout
b. 1 sacrificed ease of use by using lower
magnification
c. 1 used "lens" option - like a magnifying glass
d. 5 had strategies
1. Rapidly change size in ZoomText
2. Use Ctrl-scroll wheel on mouse
3. change window size to avoid horizontal scroll
4. Copy and paste material into Word and enlarge
5. Printer friendly version has less graphics and
is easier to zoom
a. A one column version of pages would be
much more helpful
6. Some use speech text
a. Some who would have benefited from using
JAWS did not want to move to it because
it was a statement that their condition
may deteriorate
E. About forms
1. Users who listen
a. Can't find form if its buried on the page or way on
the right
1. JAWS has a command to go to the form, but
nobody knew it existed
2. Do not know there is a form on the page until
you encounter it
3. Users want to stay in Edit mode so you can tab
from field to field
b. Can't use the form if the field labels aren't "well
behaved"
1. Need to move from hearing mode to entry mode
and back again
a. Many users had mode problems, hard to do
c. Don't hear text that is not connected to a field
1. Got to put needed information in the label of
the field
d. Have problems when pages refresh "arbitrarily"
1. Script driven check-on-entry fields
2. Time and date box causes whole page to refresh
3. When it refreshes, you are at top left again
F. Other issues
1. Using assistive software adds a huge mental load
a. Window-Eyes is more difficult than JAWS
2. Most user do not use all the available functionality
3. Many users do not know how to customize all the aspects
of their software
4. These users are just as impatient as everyone else, but
the web is critical to them
5. Separate but equal is not OK.
a. "I never trust screen reader versions because the
text version is never updated"
b. "Its double work to do text and graphic versions."
VII. We are doing it backwards
A. Hard to provide guidelines for people who magnify
B. Today, assistive technologies go on last, on top of regular
sites
1. First build site for most people
2. Then fix so site works with "special software"
C. Reverse it
1. Flexibility
a. Let people set up personal profiles
1. One column is very helpful for some people
2. Some people can handle dense information,
others go into cognitive overload
3. Wheelchairs are composed of components
4. Can swap in different components as your body
and needs change
2. Portability
a. Need ability to bring profile with them as they go
to new services, use new tools
b. Profile needs to be truly portable
3. Graceful transformation of the information to address
all of the various users diverse needs
a. CSS works on separation of content and presentation
1. http://csszengarden.com
D. How do we get there?
1. Is content management a solution?
a. CMS developers also working on understanding
information structure, granularity
2. How do we make the computer usable for everyone, break
out of the current mindset of what "accessible design
is"
3. Design your web sites so that they work well for the
client software, so that the software can be designed
whatever way the user needs
a. Real adaptation to user needs would have to happen
at browser level
b. Content should be managed in a way that enable
diverse browser functionality
4. We have to learn to relinquish our need to control how
pages appear, at the same time be able to come up with
good designs in the range of ways the content may be
presented

|- Rick Ells - 543-2875 - rells at cac.washington.edu - Rm 011S MGH Bldg -|
|- http://staff.washington.edu/rells/ -|


More information about the Accessibleweb mailing list