Re: HTML 5

28.5.2011 12:57, T.J. Crowder kirjoitti:

> Is the current practice is to comment in the mailing list or on the bug
> report?

That's what I wonder too. Both ways are supposed to work, but that's not 
really an answer.

> `b`, `i`, `u`, and `s` (strikethrough) are all currently in the
> specification[1].

All with completely rewritten semantics. Most probably, few authors will 
pay any addition to the new, rather contrived semantics and will keep 
using those elements for physical markup the old way.

This means that anyone who wishes to conform to HTML5 in old documents 
should revise _all_ use of such markup and replace it either by 
appropriate semantic markup, such as <strong>, <em>, or <cite>, or with 
the use of CSS. In _rare_ cases, preserving <b>, <i>, etc. might be the 
right move, as the intended meaning _accidentally_ coincides with the 
new definition.

That's the theory, and I think it does not have much chance of becoming 
reality. It will cause some confusion among a semantics-oriented 
minority, not much more.

Here's how I would define the <b> element, for example:

The <b> element indicates bold text. In situations where bolding cannot 
be used, user agents may ignore <b> markup or render it in a manner that 
is widely understood by users as simulating bolding. Authors should not 
use the <b> element except when quoting external sources (such as 
printed matter) containing bold text and for texts that are rendered in 
bold by convention, such as vector symbols in mathematics.

> `font` was always broken, IMHO, particularly the `size` feature. I think
> applying CSS to something semantic wherever possible, and falling back
> on an introduced span if necessary, is best.

I don't see why anyone would miss the <font> element, except in 
situations where you need to format HTML documents without CSS, and such 
situations are rare. The only situation where font would actually make 
sense is a context where the text _discusses_ fonts and shows examples 
of texts in different fonts. Then it's a matter of content, not just 
casual style preference. But this case is probably too rare to have an 
impact.

> `center` *should* have a viable CSS alternative

<Center>...</center>, being formally equivalent to <div 
align="center">...</div> but implemented inconsistently, has always been 
a problem.

The original poster had a long wishlist (the attributes align, alink, 
vlink, link, text, background, bgcolor,hspace, border, vspace, type, 
valign, width, height, cellpadding and cellspacing), so I guess he 
basically meant that presentational markup should not be obsoleted. It 
is probably too late now, though some details are still under 
consideration and some presentational markup might be "saved".

-- 
Yucca, http://www.cs.tut.fi/~jkorpela/

Received on Saturday, 28 May 2011 18:23:04 UTC