[thelist] Are you designing with CSS and web standards?

Ian Anderson ian at zstudio.co.uk
Wed Mar 16 11:01:11 CST 2005

Pringle, Ron wrote:

>For some of us, CSS isn't an option, its the rule. Part of my job
>description includes coding to current web standards (HTML 4/CSS 2), and
>accessibility is one of the primary goals. 
Brilliant - pleased to hear it! You must have a very satisfying job.

>Yes, I could have done the design in tables and it probably would have taken
>less time to develop, but it would also be much less accessible.
I'm sorry to be contrary, but "much less accessible"? I have to 
challenge that.

[caution - accessibility rant follows]

I have been specialising in web accessibility for a couple of years now, 
including nearly a year consulting with a major UK bank on their 
accessibility projects. I have tested table based designs against most 
popular screen readers and we had significant user testing with blind, 
visually impaired, dyslexic and mobility impaired users. Not one - not 
ONE - problem was encountered in any testing owing to tables. Of course, 
we designed the tables so there wouldn't be, but all the same...

The site in question was a hybrid design - simple tables with lots of 
CSS, but we had up to three or four levels of nesting in some complex 
areas. Not a peep from the screen reader users; they reported it to be a 
very accessible site. The problems they had were with higher level 
issues such as the phrasing of instructions, and so on.

Of course a site is more accessible with CSS-only layouts; PDA users 
appreciate them, I believe. But to throw statements like "much less 
accessible" around is not accurate. It's very much a marginal issue, in 
my judgement. CSS layouts are only *slightly* more accessible than good 
table layouts. Bad table layouts; well, there's no difference between a 
bad table layout and a bad CSS layout; clearly both are undesirable. 
Yes, you can have CSS-only layouts that have bad accessibility problems.

Making too much about tables and accessibility somewhat ignores the fact 
that users with disabilities have adapted to the way the web is now 
(essentially table-based) and so have their assistive technologies. 
Tables don't give current users any really big problems; if they did, 
users would ditch their assistive technology for one that could cope. 
Accessibility tools and the people who use them are pragmatic; people 
have lives to get on with, and they use the web enthusiastically. There 
are things that are extremely serious barriers to users with 
disabilities on the web, but tables generally aren't amongst those things.

If the content in the table doesn't make sense read in source order, 
that's an accessibility issue. If the content in a css layout doesn't 
make sense read in source order, that's an accessibilty issue too.

If you have a huge list of links before your content, that's an 
accessibility issue - but it's an issue with having too many links, not 
how they are positioned. You need to provide a skip link, whether the 
design uses CSS or tables. The main problem with lots of links is for 
mobility impaired users; tables or CSS positioning is entirely 
irrelevant to them. They need a skip link and that's all there is to it.

Excessive nesting *is* annoying to some screen reader users - but it 
isn't a huge access barrier.

Hybrid designs, which tend to avoid nesting with very simple layout 
tables and CSS, are *highly* accessible.

The most important issue with tables is does the page make sense when 
the content is linearised? This is exactly the same as with CSS layout. 
If the answer is yes, worry about something else, such as headings, 
avoiding excessive spans and classes and ensuring you're using good 
markup instead, grouping links together in structural containers like UL 
or OL, and so on.

To put it across as firmly as I can; CSS does not make a site more 
accessible. HTML does that. Use HTML well, tables or no tables, and you 
get an accessible web site. CSS is important for accessibility, but it 
is important because it encourages more thoughtful HTML construction, 
and discourages harmful practices; it offers nothing positive to access.

(And don't think of mentioning aural style sheets, which are a terrible 
idea, adverse to users interests, and it's a mercy no popular screen 
reader supports them.)

>Hopefully standards compliant browsers enter widespread mainstream use. I
>see IE as the single biggest factor holding back developers using CSS.
I agree, though the single biggest factor at any given moment is the one 
that's blocking your path at that moment. Yesterday it was Safari and IE 




More information about the thelist mailing list