Re: css with attribues [hardware]

On Jan 22, 2008, at 11:55 PM, David Woolley wrote:

> Brad Kemper wrote:
>
>> Unfortunately, there is an undercurrent of opinion on this list  
>> (or perhaps even a theme of the CSSWG) that authors must be  
>> protected from themselves, and that the Web must be protected from  
>> authors. It rears
>
> That shouldn't be surprising as I think that one of the main  
> reasons for introducing CSS in the first place was to remove  
> presentational attributes from HTML.

That's true. At the time, there was no other way to have distinctive,  
well-designed pages, because the "traditional wisdom" of the time was  
that any amount of design beyond the default styles was extraneous,  
and had no place in Web pages. But designers were not to be denied  
(especially not the commercially employed authors that helped make  
".com" synonymous with the Web), and they found a way anyhow, using  
the only methods they had available: multiple nested  invisible  
tables, single pixel GIFs, and a profusion of font tags.

So CSS was born to accommodate the need to design, without sullying  
the meaning of the HTML. But now, certain folk (like yourself) don't  
even seem to want design and visual style to be available from within  
CSS. In fact, in many of your postings, your goal seems to be slowing  
down, preventing, or even reversing the progress of CSS's ability to  
deal with design and presentation needs.

> I am fairly sure that the web designers of the time would have been  
> happy or even happier with just having more and more presentational  
> attributes piled onto HTML.

As someone who typically hand-codes pages, I find such pages  
reprehensible and extremely difficult to edit. I for one, welcomed  
CSS and the separation of style and content, and the eventual  
elimination of tag soup (we can hope).

I don't necessarily agree that CSS should include all the attributes  
of HTML, given that most are non-presentational. But I salute Dmity's  
creative thinking, and his ability to reason fairly cogently in a  
language that is not native to him.

There are some attributes, such as "title" or Webkit's "placeholder",  
that seem to be borderline presentational (in that they are usually  
not vital information, but offered as an aid to understanding, just  
as good design can be), and some properties, such as "content", that  
can very easily step over the border or blur the border line out of  
existence. But I think if we want to include some ideas from HTML  
attributes that we haven't already, then that should be done  
formally, on a case-by-case basis, and not to upset the existing  
ideas about the cascade, error handling, etc.

It's intriguing to me to consider the possibility of a CAS language  
(similar to CSS) to allow for the separation of any attribute from  
the HTML mark-up, into a separate shared file, but that is well  
beyond the scope of this working group, and should remain so. It  
would certainly seem like an uphill battle to get anyone to implement  
it, but if someone did so, experimentally, it would be interesting to  
see, and might yield new insights that we could all learn from.

Received on Wednesday, 23 January 2008 18:30:41 UTC