W3C home > Mailing lists > Public > public-comments-wcag20@w3.org > March 2008

Comment 2476 (and indirectly comment 2376

From: Sheena McCullagh <sheena.mccullagh@blueyonder.co.uk>
Date: Mon, 24 Mar 2008 01:27:34 -0000
To: <public-comments-WCAG20@w3.org>
Message-ID: <KNEOJNIMDPHDGIPDGJGMKECDDBAA.sheena.mccullagh@blueyonder.co.uk>
Hi,

Not sure whether this is a formal objection, some of what you have done has
made things better, but one major change, ie G 156 has made things much,
much worse and will largely hamstring dyslexics like myself and to a certain
extent others who need specific colour combinations.

Firstly, thank you for removing the requested sections, that is a great help
to those of us whom need specific colours.

Now to the problems (suggest solution right at the bottom, rationale is in
the body of this email):

Both I in 2476, and Bim Egan in 2376, were trying to explain that the only
easy way that dyslexics (myself included) and others with cognitive
impairments can get the colours we need is for the web writers to NOT
specify colours, ie G148, which has now been added into 1.4.8 (Sufficient
techniques).  I rather rambled, Bim was more succinct:

'If authors DO specify all foreground and background colors, it is virtually
impossible for users to select their preferred, or required, choice, without
also losing visual clues to menu bars etc. Colour is an important aid to
recognising menu content when reading is difficult.

Shouldn't the first sufficient technique ask authors to refrain from
specifying inessential foreground and background colors for substantial
blocks of text?'

I stated that:

'Bullet point four - this is the only one that really works,'.

Unfortunately at the time both Bim and I were commenting this bullet point,
ie 'Using a technology that has commonly-available user agents that can
change the foreground and background of blocks of text (General, future
link) ', had not been written up, it is now G 156.  At the time I did not
appreciate that this meant we would have to over-ride specified colours.  I
do not feel that anyone could possibly have known this as:

1  It is in effect virtually the same as the old bullet point one, (now the
amended bullet point 3)

2  It is not generally anything to do with web writing.  It is exceedingly
rare, in my experience, for a web writer to over-ride these controls so in
most situations control is purely down to the browser (and in the case of IE
6, the system settings) the person uses and therefore it is nothing the web
writer is 'using', as per first word of the guideline.

3  If you over-ride specified colours in FF and Netscape most Java Script
pop-up boxes and drop-down menus become unusable.  Pop-up boxes gain a
transparent back-ground, so super-imposing the text of the box on the text
of the page and the drop-down menus either go transparent or gain a
dark-grey, and unchangeable background.  I have good screen shots of this
phenomenon occurring across a range of sites, which I am sending to you via
a separate email.  (Not attaching to this one just in case it makes the file
size too large).  Ironically enough FF is the only one of the browsers that
has a built-in dictionary, a real boon for we dyslexics when filling in
forms, however it is often when we are trying to complete forms that we have
the JS issues described here.  IE doesn't have these JS issues, but doesn't
have a dictionary either.  So over-riding becomes a real no-win situation
for us.

4  It is, as I understand it, impossible to actually pass this guideline
because IE 6 works completely differently to IE 7 in relation to setting
background colours.  In effect, if you pass for IE 6, you fail for IE 7 and
vice versa.  IE 6 had a well known issue where it wouldn't over-ride
background colours in the browser unless the same background colour had also
been selected in the system settings.  This caused tremendous problems for
people needing accessibility settings when using public access computers in
cyber cafes, public libraries etc.  Some web sites managed to code their
pages so that they corrected the IE 6 problem, but in doing so, when IE 7
was released those pages now showed the problems in IE7 that had been
present on the majority of sites in IE 6.  Unfortunately in the UK vast
swathes of public access machines are still on IE 6.  It is therefore unsafe
to quote IE as allowing colours to be changed in the browser without
specifying which versions of IE will and won't.  For an example, view the
following site in both IE 6 and 7.  Try and over-ride the background colour
in the browser alone.  This was one site which corrected the IE 6 fault, but
now doesn't 'work' properly in IE 7.  For the best effect set the browser
colours to yellow text on a black background, ie the 'common' Visually
Impaired combination -
http://www.firstgroup.com/ukbus/wales/swwales/home/index.php

When I commented that this was the only one that works, I had thought you
meant a non-Java based equivalent to 'Providing a multi color selection tool
on the page for foreground and background colors', as I already knew the
information in points 1 to 4 above, it never crossed my mind that you could
possibly mean over-riding specified colours, but it appears that you do:

Examples
  a.. A Web page is designed using HTML and CSS to specify the foreground
and background colors. The user sets their own colors in Internet Explorer 7
and they can view the content with their chosen foreground and background
colors.

  b.. A Web page is designed using HTML and CSS. There is a link on the page
to instructions on how to set colors in various browsers.

This is the worst possible case scenario and completely counteracts the
benefits of G148.  Under no circumstances should G156 be used as a way to
pass this criteria.

Within G156 you have given a link to the BBC accessibility pages.  Living in
the UK and the BBC being our main TV channel, I know this site well.  I
agree that their accessibility information is second to none (but even they
don't comment on the above IE 6 problem).  However, when you actually try to
apply the settings to their web site as a whole, it becomes unusable, as it
does with large numbers (vast majority?) of web sites.

View their home page - http://www.bbc.co.uk/ without over-riding specified
colours, then over-ride them, it doesn't matter in what combination, just
the act of over-riding makes the whole page unusable, especially in FF.
Graphics disappear and writing superimposes on existing graphics to make the
text unreadable.  This phenomenon is exacerbated if you also need large size
text as they have failed to allow sufficient room for text expansion (try 32
point in FF).  We dyslexics, and others who need specific combinations, find
this again and again with web sites.  Over-ride should only be used as a
very last resort, yet you are recommending that colours are specified
meaning that over-ride would have to be used at all times.  Ironically G
148, is the only guaranteed way to pass a double A criterion, yet G156 is a
triple A.  Theoretically Triple A should be better than double A, but in
this case the converse is true.

Try this on your home page as well - http://www.w3.org/  While it doesn't
fall over in the same way that the BBC one does, all visual clues of banner
colours etc are lost.  These additional visual clues are essential to
dyslexics, something both Bim and I pointed out to you, but you
mis-understood in your reply to Bim 'We agree that it is important not to
specify some but not all of the colors.' and 'Specifying all foreground and
background color attributes of any given element in CSS', when Bim had
clearly stated that authors should refrain from this (see first quote in
this email).

You seemed to understand the same comment when I made it, but declined 'We
did not include the proposed, "Specifying foreground and background colors
of banners, features and navigation in CSS while not specifying foreground
and background colors of the main content of the page in CSS and/or HTML
(future link)" in 1.4.8. We felt that it could not be listed as a sufficient
technique because, if we are understanding it correctly, it would make it
difficult for users who need to change the foreground and background colors
for the entire page to do so.'

That comment is decidedly interesting.  If you are saying that specifying
part of a page makes changing colours difficult, what is G 156 doing in
there as a success.  It's no more difficult to change part of a page than it
is to change all of a page, you do exactly the same in the browser settings
regardless of the amount you are trying to change.

I accept that those who need specific combinations across the whole page, eg
Visually Impaired people using gold text on a black ground, would have to
over-ride if any part of the page colouring was specified and the attendant
problems that causes, eg BBC site and JS listed above.  However, I cannot
see web writers accepting that they are not allowed to specify any colours
and what's more, that creates problems for dyslexics with the loss of
attendant visual clues.

The whole point Bim and I were making is that peripheral items eg navigation
bars should have colourings specified, but leave the main text areas of the
page unspecified.  As someone who is dyslexic and has spoken to many other
dyslexics, short of a web site offering a Java type selection tool where we
can specify specific blocks, eg navigation, headers etc, the suggestion Bim
and I both made is the best possible solution.  Most dyslexics can manage
small amounts, eg navigation, in any colour combination, but need our
specific combinations to read the text body of the page, ie the main
content.  Small web site would often not have either the skill and/or
resources to provide such a selection tool, but could create perfectly
accessible, from a cognitive disability perspective, web sites by combining
the current items 2 and 3 in sufficient techniques first requirement, yet
currently these two are mutually exclusive.

The following web site is largely the model I would recommend, OK they have
made errors in that there are places where link and heading text are
specified and backgrounds aren't, but in the context I'm talking about, the
main body text and background can be changed to any colours without needing
to over-ride within the browser.  This means that we keep our additional
visual clues of coloured headings, navigation and separator bars but can
still read the body of the page.  The site is a very large one so these
additional colours are essential to readily tell what part of the site you
are in.  However, if you over-ride the specified colours, we loose these
essential additional clues (and in certain places have the JS problems see
screen shots in separate email).  Try a variety of combinations and you will
see what I mean about the ease with which colours can be set:

In FF (1.5 and 2) Tools/Options/Content/Colours - uncheck 'Use system
colours', then set the four boxes to any colours you desire.  Select OK
twice and the page will have changed colour.  Please choose at least one
combination with a background other than white.  Now do the same, but
over-riding the specified colours, ie following the BBC instructions and
look at the difference.

Please also try IE 6 and 7 again over-riding and not over-riding.  The BBC
instructions give how to over-ride, for the not over-riding test, simply
ignore the accessibility tab instructions.  In IE 6, you will see the
problem of the background not changing for the specified sections of the
page.

Here is the link -
http://www.bristol.gov.uk/ccm/content/Leisure-Culture/Libraries/invitation-t
o-save-money.en

I'm sorry this has rambled on so long, but this is a subject exceptionally
dear to my heart, it is something I live, both as a dyslexic and a web
writer.

Please, I beg you, remove G 156 and amend the sufficient techniques first
requirement to allow sites to combine items 2 and 3 as suggested above.

Sheena McCullagh
Received on Monday, 24 March 2008 03:30:09 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Sunday, 17 July 2011 06:13:25 GMT