W3C home > Mailing lists > Public > w3c-wai-gl@w3.org > April to June 2004

Re: Proposed guideline: Allow users to control the rendering of the content

From: Charles McCathieNevile <charles@w3.org>
Date: Sat, 1 May 2004 08:10:49 -0400 (EDT)
To: Joe Clark <joeclark@joeclark.org>
Cc: WAI-GL <w3c-wai-gl@w3.org>
Message-ID: <Pine.LNX.4.55.0405010736280.3547@homer.w3.org>

On Fri, 30 Apr 2004, Joe Clark wrote:

>One appreciates Yvette's suggestions. Sadly, there are issues!

>>>According to spec, pixels are defined as a relative unit of
>>>measurement, not absolute. This has clouded the issue
>Only inasmuch as the spec was ignored in writing WCAG 1.0. The anti-design
>ideology prevailed.

Hmm. Sort of... the CSS spec, If I Recall Correctly, doesn't really suggest
that pixels should be a readily variable size under user control, which is
the essential requirement for reasonable user control of rendering size.
there is also very little a user can choose as a base from which to choose
their preferences in a way that will meet some expectaion of designers -
unless we accept something like "what IE does in Windows by default" as a
baseline. In practice people often do assume 12 points as a default size,
equating to some number of pixels I forget, so maybe it is no worse than
any other arbitrary choice.

>> The biggest problem with pixels at the moment, is that Internet Explorer
>> doesn't let you scale them using the text size setting in the view menu.
>There are lots of ways around that.

Right. Many other browsers allow for proper zoom now, as well as text size
control (the two seem like different things really). But significant ones
like IE and Safari don't :-( This is an implementation problem I think.

>> Allow users to control the rendering of the content
>> Level 1 success criteria:
>> 1. If the technology supports relative measures, specify font sizes in
>> relative measures.
>I just don't see how that is necessary. I am aware that oldschool WCAG
>supporters think assigning any font size is tantamount to human sacrifice,
>but they were and are quite wrong.

[aside] Maybe you overestimate peoples' aversion to human sacrifice? (I
personally note taht there is a preparedness on the part of some to sacrifice
the effectiveness of some humans to concerns that I think are trivial, but
that kind of diversity of opinion is why we have these discussiuons :-)

see below for more on this point (of a less facetious nature).

>> 2. Make sure the content is still functional if rendering suggestions from
>> the author are not followed.
>I really don't know how that's our problem. We're writing valid code...
>apart from providing alternate stylesheets, which is absurdly easy, what
>else are we supposed to be doing?
>We're writing Web content, not presaging every oddball misrendering of
>valid content.


>> 1. Do not change the default font size of the "main text".
>"Body text" (as distinct from "display").
>> Who benefits from guideline 1.x:
>> * People with limited vision can change their font settings so they can read
>> the content in the font size they prefer.
>As has been explained several times on this list, standards-compliant Web
>developers solved this problem years ago through stylesheet-switchers.
>It's also a user-agent issue, that is, the reader's problem, not solely

It is indeed partially a user-agent problem, partially something that can and
in my opinion should be addressed by content producers offering styles which
include rendering at the users default size.

But I think there is a benefit (admittedly not at the same level as some
others) in providing a default presentation that uses the user's preferred
size. Having to change styles for each site I use is a cognitive load I don't
need, extra work that my hands don't need (and I have reasonably efficient
access to the functions offered), and doesn't transfer from site to site
easily unless it is provided  by the under-implemented but valuable
"alternative stylesheet link", which was only specified last decade, and a
stylesheet switcher built into the pages. (For an example of this that
doesn't quite actually work, see http://www.theage.com.au - a site mentioned
on IG recently who have taken some big steps towards real standards
compliance, to my joy - I read the site most days).

>> body { font-size: 0.8em; }
>> Effectively this is saying to the user "I'm just going to ignore your font
>> preferences, I want you to read the text at 80% of your prefered text size".
>Using any font size says the same thing to the user. And the user may
>override it.
>This is the way we design the Web. The CSS spec (and even the still-legal
>font element in HTML) are predicated on it.

True, but for the reasons I suggested above, it is one that does have an
impact on users. While using size variation is indeed an important cue in
typography, it seems perverse in a medium which almost uniquely offers
control over the baseline size to a user that a designer would seek to usurp
that baseline. Which is quite different from using the normal sizing features
of typography to emphasise particular features or parts of content. Despite
the relatively poorer quality of online typography compared to print,  I
regularly come across sites which think it is sensible to use a smaller type
than that I typically find in my printed reading material. If some human
sacrifice is less heinous than others, then some typography decisions are,
likewise, less user-friendly than others...

So where does that leave us? I think there is value in asking authors not to
make normal text bigger or smaller than a user likes it - but I am not yet
ready to compare this value to that of other requiremmetns we have and will
have. (Clearly it isn't the most important thing in terms of preventing

But as noted above, I think there are some issues with the proposal, and look
forward to a revised version.


Received on Saturday, 1 May 2004 08:28:36 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 16 January 2018 15:33:49 UTC