I started this as a post to CSS-discuss, but as I typed I realised that
it might be a bit off-topic for that list and that WSG might be a better
recipient. I know many people here also subscribe to CSS-D, so if you
think it would be suitable matter for that list then say so and I'll
perhaps post it there also.
Apologies for the length of this post.
-- -- --
In a thread on the CSS-Discuss list ('Accessibility + font sizing')
David posted the following:
>If accessibility is important, don't specify a font size. Leave it up to
>the visitor to be using the font size they find preferable.
This revisits a question that still really vexes me [1]. Certainly, if
the focus of the site is maximum accessibility (example: a that site
deals with disability issues) then David's advice is clearly correct,
and it could be argued that it is correct for *any* site.
However, this brings us back to the fact that for many people the
browser default text size of 16px is too large and makes the page look
ugly. I suspect (without evidence, I admit) that the reason more people
don't complain about pages with enormous type is that so many pages have
been designed to adjust the text size downward from the default -- for
example, to the relatively common 12px.
So, as a designer, I choose between two approaches:
1) 'Bottom up' approach: Design the site with smaller type (there are
plenty of methods to pick from out there) and make sure that the text
can be scaled upward (without breaking the page) by those people who
need to do so, or
2) 'Top down' approach: Adopt the policy advocated by David (and others)
of not specifying a type size at all (so the site renders using the
visitor's default type size), and allow people who do not wish to see
large type to scale downward using their browser controls.
Many in the web design community seem in favour of the Top Down approach
(no type size declared at all) but I'm not quite convinced yet and just
wanted to rekindle the debate in an effort to garner some further
insight for myself.
The following points outline my qualms regarding the Top Down approach:
1. Large numbers of users have their browser set to default text size
without even knowing it; most web users, I believe, have little idea
that their browser even /has/ a default text size setting (The exception
to this is probably those who *need* to change their text size for
reasons of accessibility -- see below). The idea that a user's default
text size is a matter of pro-active choice is therefore probably false
unless the user is part of the exception noted above.
This bring into question the advice of the W3C tips page <http://
www.w3.org/QA/Tips/font-size#goodcss> where it states:
"1em (or 100%) is equivalent to setting the font size to the user's
preference".
The above statement makes the implicit assumption that 'Browser Default'
equals 'User's Preference', an assumption that I can't help but question.
2. The majority of 'Accessibility users' (for want of a better term)
will, by contrast -- assuming that they use browsers at all -- have
their default settings tuned to their preferences, and will be
reasonably aware of how such settings are altered. Many will have a
large minimum font size specified, and/or be using IE's facility to
ignore any font size settings specified by the page.
2. As I mentioned earlier, I think it likely that 'non-accessibility
users' often *don't* see pages at default text sizes because the
designer has specified something smaller via CSS [2]. Therefore, when
they *do* visit a page with no type sizes specified (or that simply
specifies font-size: 100%) they have no idea why the page looks overly
large and clunky to them compared to the other pages they visit, nor do
they know how to scale down the text -- they simply assume that it is
designed to look that way (and, in a sense, it is).
If the designer has assumed that people who like smaller type sizes will
adjust their browser settings accordingly, he or she will probably be
disappointed much of the time.
3. Nevertheless, it is undeniably true that some people (myself
included) feel that 16px text is /slightly/ too large from a 'design
aesthetic' viewpoint [3].
4. This being the case, clearly /someone/ is going to be doing some
resizing of text when they visit your page -- whether it is the person
with perfect vision scaling things downward, or the person with
accessibility issues scaling things up. Which way it goes all depends on
where the designer pitches the 'baseline' -- Top down or Bottom Up.
5. If we use a Top Down approach then we ask people to scale downward
if they want to; but many will not know how to.
However...
If we adopt a Bottom Up approach then we ask people to scale upward if
they need to -- and many of those who need to will already have such
scaling performed automatically by their *pro-actively chosen* browser
settings. Even if that is not the case then they will probably be aware
of how to increase text size on a page-by page basis using browser controls.
6. Does this not make a stronger case for a 'Bottom Up' approach to text
sizing, where the designer specifies whatever type size he feels makes
the site look and feel the way he/she wants it to?
Designing a page with clunky 16px height type in mind feels to me like a
'lowest common denominator' approach, especially when my gut instinct
tells me that the 'corrective measures' (ie. scaling text to get the
page to a level agreeable to the specific viewer) are far less likely to
occur if I design with larger text than if I design with smaller text.
Would a Bottom Up approach not have more chance of giving everybody what
they want to see?
If you're still reading by this point, thank you.
I look forward to hearing everybody's opinions!
--
Rick Lecoat
[1] I am not talking about the merits of general accessibility coding
-- alt and title attributes, semantic (x)html, table headers/scope,
avoiding px/pt units, etc, which are all essential components of any
accessible design. I'm *just* talking about baseline type sizes.
[2] The more 'visual design' a page has applied to it, the more likely
this is to be true -- a page that simply displays the text of a research
paper often has no text size specified at all, whilst a typical
designer's page will specify text sizes in quite some detail
[3] I fully accept that this is not everybody's viewpoint. I don't even
know if it is a majority viewpoint.
*******************************************************************
List Guidelines: http://webstandardsgroup.org/mail/guidelines.cfm
Unsubscribe: http://webstandardsgroup.org/join/unsubscribe.cfm
Help: [EMAIL PROTECTED]
*******************************************************************
