Commentary on "Table Manners"

"Table Manners"?

"HTML Techniques for WCAG 2.0", W3C Working Draft, 19 November 2004, includes the statement:

That page, like many other sources, shows that accessible pages can be developed using layout tables.


These are statements in the summary at the bottom of Jim Byrne's page:

Jim Byrne's summary Response
Pages that don't use tables for layout are quicker to load.

Each cell of a layout table makes the HTML perhaps 25 bytes larger on average. The extra load-time of pages using simple layout tables will probably be undetectable. A simple layout table may be (say) a 3 x 3 table with no presentation attributes. It should work well.

A layout table with lots of rows, lots of columns, and/or lots of height & width attributes, should be consigned to the trash-can, along with the tool used to develop it.

Pages that don't use tables for layout or CSS for layout tend to be more accessible.

"Tend" is misleading. Do it right and there shouldn't be a problem. Do it wrong and there certainly will be a problem.

So - do it right! Then sleep well at night.

Pages that don't use tables for layout are more compliant with W3C standards.

What does "more compliant" mean? I am not aware of grades of W3C-compliance. It is pass or fail.

Pages using table-layout can easily be 100% W3C-compliant, even for HTML 4.01 Strict or XHTML 1.1. Pages using CSS-positioning may be non-compliant. Use "compliance", rather than "table v CSS-P", as the proper quality check.

Currently tables are more accessible than the use of CSS for layout, unless designers are prepared to do some research to get-around known and unknown bugs in current and legacy browsers.

I believe the starting-point should be "accessibility when using specialist adaptive technology", not "current and legacy browsers". Eg. IBM Home Page Reader.

The issues of accessibility are at a higher level than the standards used.

Tables should still makes sense after 'linearization'. Yes. And this is relatively easy. As Jim Thatcher says "while layout tables usually linearize well, data tables do not".
Relative units should be used rather than absolute units. Yes. But even this statement is a bit simplistic. "px" is a relative unit, but can cause problems.
Techniques to avoid repetitive navigation should be incorporated into your Web page design. Yes. This is an example of "The issues of accessibility are at a higher level than the standards used."
Use the 'summary' attribute to provide a summary of a tables content, when appropriate. Yes.
Provide an alternative single column layout when it is not possible to make a multiple column table accessible. Yes.
Use HTML tags such as TH, TD and CAPTION correctly. Yes, of course! Use all HTML correctly!
Help to make complex tables accessible by using, 'id', 'headers' 'scope', 'abbr' and 'axis' attributes. Yes. But this is not about layout tables, of course.