Re: Legacy elements (was : Complex Table Examples)

Philip & Le Khanh wrote:
> 
> David Woolley wrote:

>> In any case, if they are not in HTML, supersets of them are needed in 
>> HTML, 
> 
> Yes yes yes!  Dialects of HTML, not core HTML itself.

No.  HTML has to be able to handle all textual documents.  The allowed 
compromise is that some elements have to have a rather generic meaning 
(I exclude <i> used in the common English language uses of italics, as 
that only has the necessary generic meaning in a few human languages). 
The ability to precisely render a document type has be compromised, but 
that document type mustn't be removed from the class of documents for 
which HTML is a reasonable vehicle.

> Let the creation of dialects be so simple that each discipline
> can generate its own, without needing to burden core HTML

This is div/span HTML, which I don't believe that you want.

> with elements drawn from a vocabulary that fewer than
> 5% of web authors will ever use.

But, at least for var, I would suggest that most authors ought to use at 
various times.  I'd also suggest that at least half authors writing in a 
commercial context will need to use the kbd and sample concepts at some 
time.  It's probably only academics in humanities who would not need kbd 
and sample.  (Linguists would need the var concept, so that is in a 
different category.)
> 
>> Ah! You mean span/div/a/img/script/embed.  
> 
> I do mean "embed", but none of the others, since all the others
> feature in real life documents from every possible discipline.

Actually this was a list of the elements that the man in the street 
would probably consider as the only ones they need; it wasn't a list to 
eliminate, but the only ones to keep. (I'm in devil's advocate mode here.)

> All but "embed" have a real r\^ole to play : "embed" is superseded
> by "object", as I know you know (!),and only Microsoft's perverse#

"embed" has the role to play in terms of the elements that people 
actually use in documents.  I should probably also have included object, 
as it is needed for ActiveX, which is less of a man in the street 
feature, and is used in the sample code for embedding things like Flash. 
  In real life, embed sometimes works where object doesn't, e.g. it 
seems to be the recommended way of including SVG content.

> hijacking thereof prevents its real-life universal adoption.

Exactly, but that does affect how authors use HTML, and you are making a 
case in terms of how they actually use HTML, rather than how they should 
use it.

If you exclude semantic elements simply because few people use them when 
the context demands, one does, pretty much, end up with presentational 
HTML.  You mentioned 5% earlier.  I suspect that this close to the 
proportion of authors who think about semantics at all, especially when 
you include amateur content.

Received on Sunday, 13 May 2007 13:04:28 UTC