Re: Cleaning House

On May 4, 2007, at 12:43, Jacques Steyn wrote:

> If a user sees a *b* or *i* button on the interface, that does not  
> mean the authoring tool should write a tag in the code that says  
> *i*. If logic and presentation should be kept separate (which I  
> strongly support), the coding agent could even write
> <span style="font-style: italic">blah blah</span>.
> Programmatically this string of code and *i* would be handled  
> exactly the same.

If authors use <i> or <em>, aural UAs can easily attach presentation  
styles to these so that they are rendered differently for normal text.

If authors use font-style: italic, a non-visual UA has to examine the  
computed style for the visual media in order to determine which runs  
of text should stand out of normal text.

> On *i* and *em*: some have stated that *em* is semantic. That is  
> debatable.

Their de jure definition matters remarkably little. When tools like  
Dreamweaver emit <em> when the user presses command-i, for practical  
purposes markup consumers cannot perform any more semantic reasoning  
from <em> than they can from <i>. I'm in awe that some people seem to  
genuinely believe that changing an identifier from one string to  
another magically adds semantics.

-- 
Henri Sivonen
hsivonen@iki.fi
http://hsivonen.iki.fi/

Received on Friday, 4 May 2007 15:02:06 UTC