- From: Jason White <jasonw@ariel.ucs.unimelb.EDU.AU>
- Date: Mon, 3 Nov 1997 09:06:01 +1100 (AEDT)
- To: w3c-wai-gl@w3.org
While reading the full text of the Trace Centre's guidelines, version 8.0, the following comments came to mind. As the subject line of this message indicates, these are only first impressions, based on a quick review of the document. 1. All of the comments which I made with respect to the quick summary of the guidelines, apply to the full document as well. 2. Terminology: as noted in the HTML specification, there is an important distinction to be made between two related terms which both refer to aspects of markup, namely, "tag" and "element". The term "element", as I understand it, refers to the entire construct, including its start tag, its content, and, if relevant, its end tag. The term "tag" refers either to the start tag or end tag. The Trace guidelines consistently use the word "tag" (as in "tags affected"). This usage should be checked for correctness, and the word "element" substituted in those cases in which the demands of precision so require it. Also, the term "screen reader" is used throughout the document to refer not only to conventional screen reading software, but also to assistive technologies that are actually "aware" of the markup of the HTML document and can process style sheets. Daniel Dardailler's dichotomy of screen readers as contrasted with "markup aware" access agents or assistive technologies, is I think an important distinction, as it marks the division between the predominant solutions of today and those that will be relied upon in the future. T.V. Raman's Emacspeak is no doubt an excellent example of the latter approach. Thus, the terminology used in the guidelines could be reviewed in this respect. 3. Language and directionality: investigations should be conducted into whether HTTP headers will provide sufficient information to a "markup aware" braille or speech access product concerning the language in which each HTML document is written, or whether authors should be advised to include the HTML 4.0 LANG attribute in the header of each document. In any case, authors should be advised to use the LANG attribute whenever a phrase or passage of text within the document is written in a different language from the main body of the text. Both speech synthesizers and braille translation software need to apply language-specific rules. 4. Image maps: the last example of an accessible image map takes advantage of the OBJECT element and uses the TITLE attribute of each anchor to convey a textual alternative for each active region of the map. In this illustration however, the anchor elements have no end tags, and nor do they have any content. This seems to be incorrect. In the DTD fragment given in section 13.2 of the HTML 4.0 specification, both the start and end tags of the A element are marked as required. Also, the OBJECT element enables quite sophisticated long descriptions, containing embedded links, to be written as the content of an OBJECT. It was this flexibility which persuaded the WAI working group to argue persistently, and ultimately with success, for the retention of the SHAPES attribute in OBJECT and the COORDS attribute of the anchor element. The guidelines should reflect the full potential of OBJECT to provide fully formatted long descriptions, rather than just a list of links such as can be created by the ALT attribute of AREA. 5. Once the final HTML 4.0 specification has been released, all references in the guidelines to specific HTML elements should be linked to the corresponding definitions in the specification. 6. In general, it is important to emphasise that the guidelines are excellent as they stand, and it is evident that much effort has been expeded in their development and maintenance.
Received on Sunday, 2 November 1997 17:06:24 UTC