RE: Linear reading order

John M Slatin writes:
 > 
 > But the proposed wording still isn't testable: the problem is with the
 > word "logical."
 > 
 > Perhaps a case could be made that Yvette's example of the sentence
 > broken into paragraphs that are floated left and right violates 3.1
 > because the order in which the phrases occur in the default
 > presentation-- the source document as it would be rendered by user
 > agents that don't support style sheets-- breaks so sharply from the
 > principles of English  grammar and syntax that meaning cannot be
 > determined

Yes, and this is also true for user agents that do support a style
language, but which don't apply the author's style declarations.
"Screen" is not the only CSS media type, and CSS is not the only style
language.

Yvette's example violates the text and purpose of guideline 1.3, but
not, as currently worded, the success criteria. Perhaps what we need
is:

Structure can be derived programmatically without analysing the layout
or presentation of the content. (this would be in guideline 1.3 level
2).

The structure of Yvette's example is a single paragraph comprising one
sentence. Discounting the style attributes, this structure cannot be
derived programmatically because the markup specifies two paragraphs
ad the order is wrong.

Note that my proposal is about separation of structure from
presentation, not about reading order as such. I don't think reading
order needs to be discussed separately except perhaps in an example,
but not in the success criteria.

Received on Wednesday, 7 July 2004 21:04:05 UTC