- From: Jason White <jasonw@ariel.ucs.unimelb.EDU.AU>
- Date: Wed, 29 Oct 1997 09:35:54 +1100 (AEDT)
- To: WAI HC Working Group <w3c-wai-hc@w3.org>
I have now read some of the messages which Dave has helpfully cited, though not all of them, due to a local network problem. However, I have several comments to make. 1. The Q element is important for purposes of accessibility, since it allows audio and braille styles to control the rendering of quotations. This functionality can not be achieved if the quotation marks are inserted directly. 2. In respect of audio user agents, the quotation marks are irrelevant. Thus, regardless of whether or not it is decided that the HTML 4.0 specification should adhere strictly to RFC 2070, it is obvious that speech-based user agents will employ changes in voice characteristics rather than quotation marks to highlight in-line quotations. To that extent, they constitute an exception to the user agent behaviour mandated by RFC 2070. 3. In English braille, the conventions concerning quotation marks are slightly different from those of print. Double quotation marks are used to denote the outermost level of quotation; inner quotations are indicated by single quotation marks. This is the inverse of print conventions. Thus, when formatting a document in braille, it would be preferable that the insertion of quotation marks be the responsibility of the user agent, subject to any applicable style sheets. 4. It is the responsibility of the HTML working group to decide whether the specification should conform to RFC 2070 in requiring the user agent to introduce the quotation marks. From the standpoint of accessibility, the most important point is that the Q element be implemented by user agents and deployed widely by authors in their HTML documents. Whether compliance with RFC 2070 would constitute a significant obstacle to its widespread acceptance, and if so whether this justifies a departure from the RFC requirement, is an issue that is best resolved by the HTML WG in negotiation with vendors of user agents, authoring tools, etc. My personal opinion is that by default, user agents should detect whether quotation marks have already been supplied at the beginning and end of the quoted passage, and insert them only if they are not already present. A superfluous set of quotation marks is as undesirable in braille as it is in print.
Received on Tuesday, 28 October 1997 18:07:36 UTC