Re: Table Techniques - Summary

Does the subject line mean "summary of table techniques" or "table
techniques for the summary attribute"? Quick, LD-boosters-- make subject
lines like this illegal on the Web!

> Layout tables must not have a summary (not even a NULL summary). This
> reverses our earlier decision that layout tables may have a summary.
>
> The rationale behind the no summary rule is:
> - layout tables should not be used (use CSS)

As long as HTML4 and XHTML 1.0 are W3C recommendations, people can use
tables for layout if they want. You don't have to like them, but they *are
not prohibited for layout*. The spec merely says "Tables should not be
used purely as a means to lay[ ]out document content as this may present
problems when rendering to non-visual media." "Should not" does not mean
"must not," and I challenge advocates of all-CSS-all-the-time design to
provide a list of current "non-visual media" that cannot handle layout
tables.

<http://www.w3.org/TR/REC-html40/struct/tables.html#h-11.1>

CSS layouts are better in myriad ways and tremendously desirable, but
tables are not prohibited, nor should they be. WAI is not a higher power
than the W3C's own HTML working group.

> - it appears that layout tables will be deprecated in XHTML2

XHTML2 is even farther away from ratification than WCAG 2 and has no
bearing on our deliberations here.

> - the function of the layout table summary can be better expressed elsewhere

Again with the misconception that anyone gives a damn that the table, an
underlying HTML structure invisible to the visitor, is used for layout.
You don't need to "express" the "function of the... summary" at all. Just
use your table for layout. Don't get all meta on us.

> - we should not require a NULL summary just to make the author "jump through
> hoops"

That isn't much a hoop to jump through. This esteemed working group
separately and elsewhere tentatively proposes that authors must police how
many nouns come in a row and add vowel markings to Hebrew, as though we
were children or something. summary="" is *nothing*. It's trivial.

summary="" and an absent summary are so similar that the former should not
be prohibited.

> Data tables must have a valid summary.

Why every single data table? What if it's a 2x2-cell table that perfectly
explicates itself? Why "make the author 'jump through hoops'"?

<http://www.w3.org/TR/REC-html40/struct/tables.html#adef-summary>

summary is an *optional* attribute in HTML. If you leave it out, your
table will still validate. Explain how WAI can require an attribute that
the W3C (X)HTML spec itself does not require.

> Proposed description of a valid summary:
> The summary must describe the relationship between cells.

The myriad table *headers* do that. Yet again the WAI WCAG WG (take your
pick) wants the entire Web turned into words words words, except inasmuch
as words are prohibited or tightly regulated because a learning-disabled
person may find them confusing at some unspecified future date.

WAI (sic) has a poor record in explicating the true requirements for table
accessibility, and this is not making things better.

> The summary does not have a maximum length.

The (X)HTML spec doesn't say there is one, so the statement above is
meaningless.

> The summary must not contain placeholder text.

True. Not even a space character.

> If the summary is less than 20 characters then it is suspicious.

That's very amusing. You realize how compact Chinese and Japanese can be,
right?

Here we go again with telling people how to write.

> Can the summary just link to another document (kinda like a longdesc)?

No. Like alt, it cannot contain markup. WAI WCAG WG (or whoever) should
know that already, but then again, considerable ignorance of the HTML spec
has already been demonstrated.

> What if the table is summarized in the document - do you still require a
> summary attribute?

No.

Glad to see this has all reached the stage of "resolution." Too bad it
stinks.

--

  Joe Clark  |  joeclark@joeclark.org
  Author, _Building Accessible Websites_
  <http://joeclark.org/access/> | <http://joeclark.org/book/>

Received on Thursday, 7 August 2003 14:51:47 UTC