W3C home > Mailing lists > Public > w3c-wai-gl@w3.org > July to September 1999

RE: Checkpoint 3.3

From: Wendy A Chisholm <chisholm@trace.wisc.edu>
Date: Tue, 13 Jul 1999 09:48:15 -0500
Message-Id: <199907131453.JAA46170@mail1.doit.wisc.edu>
To: w3c-wai-gl@w3.org
Let's look at three checkpoints together:  

3.3 Use style sheets to control layout and presentation. [Priority 2] 
5.3 Do not use tables for layout unless the table makes sense when
linearized. Otherwise, if the table does not make sense, provide an
alternative equivalent (which may be a linearized version). [Priority 2] 
                   Note. Once user agents support style sheet positioning,
tables should not be used for layout.
                   Refer also to checkpoint 3.3. 
5.4 If a table is used for layout, do not use any structural markup for the
purpose of visual formatting.  [Priority 2] 

3.3 should have a note that says, refer to 5.3 so that it is more obvious
that we are saying tables are all right to use until style sheets are more
widely supported.  Perhaps an "until user agents" clause would also be
appropriate, not sure.

Despite that this is not highlighted, it DOES say that tables may be used
for layout until style sheets are more widely supported.

Here at the Trace Center, we also have been working to redesign our site so
that it is more hip as well as follows all of the guidelines (as well as
creating examples as the guidelines have developed).  There are many
interesting issues with style sheets, but as Jason said, they do transform
gracefully.  We also get *many* hits from Netscape 2 browsers - last month
we had over 16,000. so we are constantly testing with a *wide* variety of
browsers on a variety of platforms (side note:  it's amazing how different
the implementation is of a single product for Windows versus Macintosh!
not only in display but features as well!!)

Today, it seems that the best option is to use browser sniffing and serve a
style sheet based on browser type - if you want to use the CSS2 layout
features.  Although, the number of authors who know JavaScript is probably
less than those familiar with Style Sheets.  HOWEVER, the script is very
easy to personalize for a site and we could write up a set of instructions
for its use (cut, paste, edit the x, y, z variables, and post to your site).

We have found that style sheets work best on the extreme ends of the
spectrum.  for example, they transform great for 2.0, 4.0, and text-based
browsers, and not so great for the 3.0 browsers - but we've made it work
(there are several tricks to use).  We have not run across a situation
(that I can remember) where the use of a style sheet has ever made
something unusable, only not as perfect to look at.  for example, we wanted
to use a list with a background color.  It worked perfectly in IEx, but
there was a white border in netscape 4.  This is a designer's nightmare
because they aren't able to force the same design on everyone, but it is
still accessible, usable, and most likely the user has seen it on other
pages using similar techniques.

There have been issues with turning off style sheets in IE, but I believe
we've found work arounds for those as well.

Robert, perhaps part of your education and outreach should include "how to
use style sheets that transform gracefully."  It sounds like you are
dealing with several issues (as we all are):
1.  Authors want their legacy pages to be AA compliant without much work
(lots of tables for layout).
2.  Authors don't have time to learn the newest technologies and the tools
aren't helping them.
3.  Older browsers are still widely used and they don't support the latest
4.  Designers are unable to accept that they can not "force" their view of
the information onto everyone.

Looking at these issues in more detail:
1. - We say that tables are all right.  The working group was very aware of
legacy pages and the extensive use of tables.  We just strongly suggest
that you move to style sheets when the time is right for your organization.

2. - Not much we can do other than give lots of examples, point them to
tools that do work, and hope for the best.

3. -  As I said before, people with older browser are probably used to
seeing things "broken."  The guidelines are meant to help *guide* designers
and authors to use techniques that will allow pages to be usable and
accessible on older browsers.  We don't guarantee that they will look as
pretty, but how can they when they don't have the same features as the
newer browsers?

4. - Designers will have to let go of the need to make pages look the same
in every viewer.  I attended a keynote "conversation" at Web99 between Don
Norman and Jakob Nielsen a couple weeks ago.  They were trying to get this
idea across.  Especially as people begin accessing the Internet with all
types of appliances, many of which either do not have a visual display or a
very small one.  

I also believe a certain amount of education and outreach needs to be
directed at users (to let them know that better tools are available) and to
browser developers.  I am happy to see some of the campaigns that have
sprung up asking browser developers to follow specifications.  Authors are
getting fed up, as well they should.  Web documents were supposed to be
portable, but the browsers have broken that.  I anxiously await the
implementations of the UA and AU guidelines as well.  We can't ask content
authors and designers to do all the work, it needs to be a joint effort.
Many things are broken, let's try to get them fixed.  (that's the naive
idealist speaking - here to change the world! <grin>)

I'm not sure if this helps at all.  It is clear that the working group
needs to make our assumptions clear, to interpret some of the abiguities,
and to continue fleshing out examples in the techqniques document.  This is
the work that lies ahead, I hope everyone is ready to participate!


At 06:37 AM 7/13/99 , you wrote:
>You said the magic words in your response, "why style sheets should be
>used".  Then the guidelines should reflect "Should"
>Jason said >Support for CSS has existed in major browsers for at least the
>past two years, and though it is inconsistent, it is steadily improving.
>rob>it is not there yet across the board! Do you want me to build an online
>catlag using CSS and have customoers come to it and watch it blow up or not
>be able to use it in older browsers - because it has problems degrading

>gracefully, because something was added at the last minute and not checked
>on Netscape 2. Yes we still get hit by people who use older browsers with
>multiple versions.  There are people who do not upgrade and they are our
>customers.  Do you want me to provide information to people and have then
>not be able to access it or at least have difficulty?  Both of these
>scenarios will have us answering a congressional response - not fun!.  If
>this is the environment that you want all the federal web developers to live
>in, then you may not get much support.  Most web develoeprs are just now
>understanding HTML 3.2!
>I am trying my  best to Educate and Outreach to plant seeds on what is
>coming and to let them hear this for the first time.  Most people do not
>know that PDF can be posted as long as there is a HTML version.  Most people
>have no ideas what CSS is.  We have a lot of education to do!
>jason>due to the importance of style sheets as the only technology which
>supports rich visual (or auditory)presentation, with retention of the
>document's logical structure and markup semantics.
>rob>this means that every government office will be required to switch to
>CSS.  This is not going to happen and you will be alienating the web
>developers that you want on your side.  They have no money for training and
>I see no plans where funds will be allocated by the Rehab Act Section 508 to
>provide for training.  This environemnt is not client-server, it is
>everyone, technical and many non-technical.  
>They can still use HTML 4, tables for layout and without depractated fonts.
>Get them to this point and give the web developers time to adopt CSS and
>hopefully someone will detail when CSS blows up in different versions of
>Netscape.  But unfortunately tables are a big part of life in the government
>and CSS is not.
>Now, we am implementing CSS on our intranet because we use one browser, but
>we are lucky.  Most offices are a kludge and the politics are such that the
>web developers has no control of the environment.
>-----Original Message-----
>From: Jason White [mailto:jasonw@ariel.ucs.unimelb.EDU.AU]
>Sent: Monday, July 12, 1999 7:26 PM
>To: Web Content Accessibility Guidelines
>Subject: Re: Checkpoint 3.3
>Without using style sheets, and without confusing the distinction between
>structure and presentation, there are few means available in current HTML
>technology by which to control layout and presentation while maintaining
>correct document structure. Visual presentation is, however, important;
>and this is why style sheets should be used. The guidelines are careful to

>require that only style language features supported by user agents be
>employed. Support for CSS has existed in major browsers for at least the
>past two years, and though it is inconsistent, it is steadily improving.
>I would strongly oppose any attempt to remove or otherwise erode the
>requirement specified in checkpoint 3.3, due to the importance of style
>sheets as the only technology which supports rich visual (or auditory)
>presentation, with retention of the document's logical structure and
>markup semantics.
>The simplest solution to the practical problem would be to require
>existing web sites to be repaired up to level A conformance, whereas new
>web sites (or existing web sites when their content is substantially
>updated) must achieve double-A conformance. This is an issue that should
>be addressed to policy developers; the guidelines must conform to their
>own goals and definitions, and should not be unduly influenced by whatever
>problems may arise for developers if government policies are formulated
>which do not adequately address the issue of compliance costs.
Received on Tuesday, 13 July 1999 10:53:35 UTC

This archive was generated by hypermail 2.4.0 : Thursday, 24 March 2022 21:07:16 UTC