W3C home > Mailing lists > Public > w3c-wai-ig@w3.org > July to September 2004

RE: HTML Table Markup

From: Bailey, Bruce <Bruce.Bailey@ed.gov>
Date: Fri, 10 Sep 2004 14:13:47 -0400
Message-ID: <CCDBDCBFA650F74AA88830D4BACDBAB504FBD9@wdcrobe2m02.ed.gov>
To: "Chris Ridpath" <chris.ridpath@utoronto.ca>, <w3c-wai-ig@w3.org>

> This suggests that we could relax the TH rule to exclude small data tables.

This leads us to the slippery slope of defining "small".
I first gave the 2x2 example of two names and two phone numbers.
The consensus was:  Well, you don't need to add a header row for that.
I then expanded this to a 3x2 example of names and numbers.
Others expanded this to even 3x3 and concluded:  Well, you don't need to add a header row for that.
To which, I much respond:  How small is small?
Does "small" reflect the number of data cells or structure?  (The example[*] in the techniques document only has eight data cells.)
Is it not really a matter of how simple the table is?
How about 4x3?  If 3x3 is easy, it hardly seems like adding n-number of rows makes the data table hard!
This returns to the original example from the question posted to the Trace 508 list of n-rows of name, address, city, state, zip:  Nx5.  This is still, I would argue, a very simple table, even though it is no longer small.  It is also (per our discussion) reasonably accessible and the lack of a header row is not a significant barrier.  The table would benefit from the addition of a header row, but clearly this is more a usability issue.

P1 guidelines address obstacles and not mere conveniences.

Unfortunately, I have now worked us into more trouble as we have "very simple" data tables and "regular simple" data tables (that really do need header rows and columns, but don't have two or more logical levels of row or column headers).
Constructing a definition to discriminate between this "very simple" and "regular simple" would be quite tough.
If we only exempt "small" data tables, nothing is done to clarify application of 5.1 to the specific example that prompted this thread!

> Do small data tables require header markup (could be TH or other)?

This question still leaves us parsing "small" and "require", and I don't want to substitute "really simple" for "small"!  I think the question should be:

Is 5.1 meant to specify the addition of header columns or rows if said would otherwise be absent? 

What is the formal mechanism for achieving consensus?  Maybe we need to pursue that mode of resolution?

> If we required header markup (TH or other) for small data tables would
> that make them more accessible?

Yes, small data tables become more accessible with row or column headers (or at the very least, do not become less accessible).  "Really simple" data tables would likewise benefit.  Requiring the _addition_ of row or column headers would resolve some ambiguity and is not entirely unreasonable, but it is more intrusive than any other P1 guideline.  That might be okay.


* http://www.w3.org/TR/WCAG10-HTML-TECHS/#identifying-table-rows-columns
Received on Friday, 10 September 2004 18:13:54 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 5 February 2014 23:39:44 UTC