W3C home > Mailing lists > Public > w3c-wai-gl@w3.org > July to September 1998

RE: Can we really deprecate tables?

From: Alan J. Flavell <flavell@a5.ph.gla.ac.uk>
Date: Tue, 11 Aug 1998 12:24:04 +0100 (BST)
To: "Dobson, James" <JDobson@rnib.org.uk>
cc: w3c-wai-gl@w3.org
Message-ID: <Pine.OSF.3.96.980811112824.7889A-100000@a5.ph.gla.ac.uk>

Excuse me, but there seems to be some misunderstanding between us
here.  Let me express directly the point that I believe I am
addressing here, just to be sure that we aren't arguing along
different lines. 

As I understood it, you were saying that the migration from "tables
for layout" to style sheet layout control should be deferred until
fully CSS-capable browsers were available to _disabled_ users (in
this context we're thinking specifically of visual impairment.)

I don't see why this is either necessary or appropriate.  It seems to
me that the style sheets that are going to be supplied by the vast
majority of web page authors will be targetted at (excuse me) "normal"
users and browsing situations, and this presentation is more likely to
get in the way of accessibility to unusual browsing situations than to
help it.  For the most part I believe that readers in unusual browsing
situations will want to be turning off or overriding the presentation
suggestions in the supplied style sheet, and imposing their own
presentation.  This can be done adequately with even quite old
browsers, via their configuration options (even better in browsers
like Opera, that offer the reader a wide choice of fonts, colours
and magnifications).

I have carefully read your reply and I honestly still am unable to
understand why you consider that this migration needs to wait for
_visually impaired_ readers to have style-sheet-capable browsers. 
Quite the opposite:  your argument reinforces the idea that immediate
migration is especially advantageous to visually impaired readers, and
the reason that the migration is being resisted by many is the fact
that it's potentially deleterious to _mainstream_ browsing situations
that don't yet support CSS. 

On Tue, 11 Aug 1998, Dobson, James wrote:

> A development on RNIB's website allows visually impaired user's to specify a
> style sheet to access RNIB's webpage in there own font, colour and size etc.

I don't for a moment want to say anything against that, but surely
visually impaired readers want to access the whole range of pages that
are on the WWW, they don't want to be restricted to the ones that have
been especially tailored for their needs.

> As you may well know the term visually impaired covers a lot of eye
> impairments, and each person may see slightly more or less. This means that
> allowing the user to have as much control over what is displayed is very
> important. 

Certainly: that was my whole point.  And that's the reason why an
author-provided style sheet, no matter how carefully designed, and no
matter how beneficial for _some_ readers, is inevitably going to make
matters worse for some subset of readers who have different

> We have all seen the different ways Netscape and Internet Explorer handle
> CSS.

If it's a problem, then turn it off.

> I was not aware that you can "turn off" CSS in these browsers (tell me
> different!!),

I am frankly astonished that you are unaware of this.  
In CSS1, for example, it states:

  This strategy gives author's style sheets considerably higher weight
  than those of the reader.  It is therefore important that the reader
  has the ability to turn off the influence of a certain style sheet,
  e.g. through a pull-down menu.

I think that's tantamount to a mandatory requirement on CSS-capable
browsers to offer the ability to turn style sheets off.

As another mail has just pointed out, there are some problems with
turning it off in MSIE4, but that seems to mean that having a non-CSS
capable (e.g "old")  browser can be beneficial in some situations;
that doesn't seem to me to support your thesis that migrating from
table-for-layout to CSS needs to wait for CSS-capable browsers (if
anything, it supports the opposite conclusion, no?). 

> I have noticed that you can specify a user CSS file in IE

This could be an additional convenience, but I don't see how it is a
necessary _before_ promoting a migration from tables-for-layout to CSS

> When using CSS for structural presentation, if a browser ignores the CSS
> information it will display it all as one piece of information after
> another. If the HTML file was not designed properly then you could have
> information in locations that are not appropriate. 

CSS is not intended to be used for conveying structural information,
but for suggesting an appropriate presentation of the content and
structure: but the suggestion can only be appropriate for a certain
range of presentation situations, and is going to be useless or even
counter-productive some other ranges of presentation situation (and,
as I said before, this relates to shortcomings of browser platforms
just as much as it relates to physical impairments that the reader
might have). 

A properly built HTML/CSS document will still present its content when
the CSS is ignored, as indeed it must be in some situations (colour
information will be useless in a monochrome browsing situation, and
font suggestions cannot usefully be implemented in a speaking browser,
to take just two examples).

> This would be down to
> good design, but surely a browser can be programmed with the ability to
> ignore or change table tagging if the user requires it?

This seems to be suggesting a hack to work around the shortcomings of
another hack, when a properly engineered solution is already there.

I think I better shut up now and let other people have their say.

best regards
Received on Tuesday, 11 August 1998 07:23:45 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 16 January 2018 15:33:27 UTC