[whatwg] <img> element comments

[Michel Fortin]

Le 5 nov. 2006 à 7:42, Elliotte Harold a écrit :

> There are always edge cases. The distinction between semantics and  
> presentation is a fuzzy one. Nonetheless, I think most of the time  
> height and width as specified on today's img tags are clearly  
> presentational.

I'm beginning to get lost in this debate. Style and presentation (the  
words) applies to so many thing, but I wonder if it hasn't become a  
synonym for "visual style" and "visual presentation" in this  
discussion. I mean, by style you could mean style of writing, by  
presentation you may talk about how the author chose his words to  
present an idea; surely no one wants to see that part of a CSS  
stylesheet.

As Elliotte said, the line between semantics (the content) and what  
is the formatting (visual presentation) is sometime blurry. To me,  
what should be part of HTML is what describes the content of a  
document; whether it's textual content, visual content, or audio  
content should have no importance: if something is part of the  
content then it should be in the markup, not in an external stylesheet.

I think that ideally, such markup should express what the author  
means (<em> more semantic than <i>), but that's not always a workable  
solution, and in those cases it's good to have fallback solutions.  
The style attribute is one of them, although style="" has its  
shortcoming: it requires CSS support.

So, to what extent should HTML be tied to CSS for an acceptable  
rendering of documents? I mean, if a browser that cannot handle CSS  
cannot display the content in a readable manner, aren't we breaking  
the separation of the content from the (visual) presentation?

  - - -

Taking this back to the subject of height and width of images, I'd  
simply ask myself if the attributes are necessary to have an  
acceptable presentation. I see two cases in which the attributes have  
a benefit:

1.  while the document is loading;
2.  if the image needs to be resized.

I won't rehash case one, I think it's clear that many people consider  
that useful, useful enough at least to add these attributes to a lot  
of images even if they aren't resizing anything.

Two is tricky too. Let's say I have a high resolution image of a  
mathematic formula inline in the text (presumably to have good  
results when printing). It's clear that you don't want the image of  
the formula to be too big: in this case the resizing, while clearly  
presentational, is absolutely necessary to have a readable document.  
I take note however that it'd be more useful to set the size in em or  
in ex than in pixels, the downside is that you need CSS for that. I  
also admit that this is a made up hypothetical case; maybe no one  
does that, I've no idea.


Michel Fortin
michel.fortin at michelf.com
http://www.michelf.com/