Markup "semantics" or structure? (was: Displaying multiple copies of the same div by css-coding)

Headless <invalid_address (AT) dna (DOT) ie> wrote in reply to Christoph Paeper
<crissov (AT) gmx (DOT) net>:

Semantically <span>this</span> is 100% identical to:
Semantically <div>this</div> is 100% identical to:
Semantically this is 100% identical to:

Ah yes, "semantics," the false god of the CIWAHians and other kooks
who don't understand that one marks up *structure*.

No content structure whatsoever is implied or achieved
by span and div elements, styled or not styled.

I count grouping, even generic, as a part of structure.

I agree. Which is why elements with only one child are so bodgey:

<div>
<div>
<div>
An enigma trapped in a riddle within a paradox...
</div>
</div>
</div>

div div div:after {
content: "Or whatever the phrase is.";
}

IBTD. Marking up text (or in the wider meaning: content) and adding links
is what HTML is all about. An element that doesn't mark up anything isn't
mark-up.

img ... alt="">,

Yes, a misconcept that's finally being taken care of with XHTML2. The more
appropriate 'object' element is however already in HTML4.

Since an <img> is, by definition, a *replaced* element it makes no
difference whatsoever what its internal structure (or lack thereof in
the case of empty elements) is.

It could be said that the appropriate way to insert images is with
entity references.

The OBJECT element *type* (unless you believe there is only one
instance in the world, and we all have to share) may be "already in
HTML4" but how many HTML processors do anything useful with it. Is
including in a specification detailed description of supposed
implementation while also contradicting oneself in such a way that the
described implementation is *impossible* a sensible thing to do? I'm
thinking here specifically of the supposed cascading of nested objects
to find one the browser can display, but there are surely other
examples.

Raggett's descriptions of his fantasy "typical browser" are the proof
that HTML 4 is not really concerned with "semantics" or even
structure, but an attempt to control what I see on my monitor. A
proper specification of structural, and especially "semantic" markup
would not include any mention of implementation whatsoever.
Implementation is the role of CSS.

object type="image/jpeg" data="decorative-image.jpg"></object

Doesn't mark up anything either. Yet the above is the correct way to
insert a decorative image (enclosing a text alternative in the object
element would be wrong).

No, the correct way to insert a decorative image is with CSS. Marking
up a decoration is wrong. Where you not just a few paragraphs earlier
waffling about "semantics"?

Headless

I suppose it's hard to think clearly. :-)

--
semantics:after {
counter-increment: frownies;
}