Re: Fw: Disturbing trend in tables

On Fri, Jan 12, 2001 at 09:58:36AM -0800, Ben Canning wrote:

> Not sure I understand the complaint about  's. If a paragraph has no
> content, the browser doesn't draw it. So if the user hits enter 3 times, we
> need to at least put in   or the paragraph will just disappear at
> browse time. You could make the argument that an author shouldn't use empty

  ... and it'll disappear at browse time anyway. What is a P with only a 
  non-breaking space inside it - a paragraph in which two (non existing)
  words are not to be visually presented with a newline between them. An UA
  is more than allowed to simply not render anything at all there.

  If a user hits enter three times, the user has three times entered the
  CR character, which is defined as 'whitespace', and to quote the HTML 4.01
  specs:

   "User agents should interpret attribute values as follows: ...
      Replace each carriage return or tab with a single space."

 
  And, of course:

   "User agents should ignore empty P elements."

  A P with only an   ... well,   isn't strictly whitespace, but
  that debate has been ongoing for a while. Personally I'd say a P with an
    is as empty as a politicians head.



> <p>'s as spacer's, but that's not something that FP can enforce. Do empty
> paragraphs cause problems for screen readers?

  I don't know ... but *if* a screen reader attempts to be 'smart' and make
  a small pause before each paragraph, it certainly is gonna sound funny. I
  don't really wanna think about how it might end up on a Braille reader, but
  it *is* a structural problem.

  And, to be honest, it does give a nice example of why a HTML editor cannot
  be WYSIWYG - and why the concept needs to be thought over once more. Perhaps
  an editor based on objects would be more suited; but then the user would have
  to explicitly say "end paragraph now" ... hm.


-- 
 - Tina

Received on Saturday, 13 January 2001 16:09:04 UTC