- From: Jason White <jasonw@ariel.ucs.unimelb.EDU.AU>
- Date: Thu, 9 Oct 1997 09:45:51 +1000 (AEST)
- To: WAI HC Working Group <w3c-wai-hc@w3.org>
I wish to thank Al for clarifying his concept of "keys", thereby confirming my initial understanding of his intention. The main question which emerges from this discussion is whether "key" relationships, namely semantic dependencies as between data cells, need to be explicitly indicated in order to generate a fully satisfying audio rendering. Can one think of cases in which such semantic content would facilitate the reading of the table? One possibility is that it could influence the reading order, since obviously a cell on which related data depend must be read prior to the latter; but such a relationship is normally indicated visually by the fact that the dependent data appear in columns that are located to the right of the column on which they rely. For example, the data referring to each senator are placed in the second and subsequent columns of the sample table mentioned in the HTML draft, with the senator's name, to which later columns refer, appearing in the first column. Now a straightforward reading of the data across the rows would adequately convey such a relationship, which can readily be inferred by the listener from the meaning of the text. Thus, there is a need for examples of circumstances in which the encoding of "keys" relationships would make a significant difference to the comprehensibility of the audio rendering.
Received on Wednesday, 8 October 1997 19:46:13 UTC