On Sep 9, 2007, at 13:43, Anne van Kesteren wrote: > Although maybe you could add an additional requirement that a > header cell for a given cell needs to span the same columns (or > more) for it to become a header cell for that cell (also, not same > amount of columns, the same columns). This would also help with the > Thesis example from Henri. That's not quite right. You could reasonably have a cell spanning two columns and both columns having single-column header cells. In that case, it is quite reasonable to associate the wide cell with both. I suggest experimenting with a rule that a cell doesn't get associated with any column headers if the cell spans a column that has no header. (And analogously for rows.) Consider: <table> <tr><th>Label 1<td>Data 1 <tr><td colspan=2>Headerless cell <tr><th>Label 2<td>Data 2 <table> vs. <table> <tr><th>Label 1<th>Label 2 <tr><td colspan=2>Data 1a and Data 2a <tr><td>Data 1b<td>Data 2b <table> For reference, the "Thesis example" Anne is referring to is http://hsivonen.iki.fi/thesis/html5-conformance-checker#abstract -- Henri Sivonen hsivonen@iki.fi http://hsivonen.iki.fi/Received on Monday, 10 September 2007 08:55:12 GMT
This archive was generated by hypermail 2.2.0+W3C-0.50 : Wednesday, 9 May 2012 00:16:07 GMT