- From: Jason White <jasonw@ariel.ucs.unimelb.edu.au>
- Date: Thu, 15 Apr 1999 12:14:26 +1000 (AEST)
- To: WAI Markup Guidelines <w3c-wai-gl@w3.org>
In a further attempt to clarify my position, I would argue that checkpoint 5.3, at priority level 2, should prohibit the use of tables for layout, except where either of the following conditions is satisfied: 1. There is an alternative version of the document provided, which is properly marked up in such a way as to reflect its logical structure and to comply with a formal grammar such as a DTD; this alternative version should (i) contain the same information as the original; (ii) be updated simultaneously with the latter; and (iii) be linked to in a prominent fashion from the version that uses tables. These requirements are already provided in the checkpoint related to alternative pages and can thus be incorporated by reference. 2. (I) All of the logical structure and semantic distinctions conveyed by the visual format of the document which uses tables for layout, are also present in structural markup which is included within the tables themselves (E.G. lists and paragraphs within TD elements), such that deletion of the table-related markup would yield a valid (formal grammar compliant) document in which the aforementioned structural and semantic distinctions are preserved in virtue of the structural markup; and (II) When formatted by a visual user agent, the document which uses tables for layout appears as a single column of text. This is rather a complex way of stating the requirements but I am trying to be as clear as possible. The essential idea underlying item (2) is that the structural markup (excluding the table-related tags, which are purely presentational) adequately conveys the logical organisation and semantic distinctions inherent in the document, so that the table-related markup can be stripped and the resulting document can then be formatted with style sheets etc., for non-visual presentation. The "single colun" requirement reflects the continued existence of primitive screen readers that are unable to handle columnar material.
Received on Wednesday, 14 April 1999 22:14:34 UTC