[whatwg] Deprecating <small> , <b>

Smylers Smylers at stripey.com
Tue Nov 25 00:41:01 PST 2008


Pentasis writes:

> [Asbjørn Ulsberg writes:]
> 
> > However, as you write and as HTML5 defines it, there is nothing
> > wrong with <small> per se, and I agree that as an element indicating
> > "smallprint", it works just fine.
> >
> > Since my initial reply might have been a bit too colored by the HTML4  
> > definition of the element and its current usage on the web, I hereby  
> > withdraw my comment and conclude that I mostly agree with you. :-)

Yay, consensus!  Thanks, Asbjørn.

> But isn't this just the reason why it should be dis-used?  The HTML4
> spec defined it as a styling tag, and that is how it is *mostly* used
> and understood by the majority of the users/authors.

That may be true (though authors who want smaller text just because they
think the default looks too large could also use <font size=2> or CSS),
but authors who wanted to diminished the emphasis of certain content to
users are likely to pick <small> because there isn't much else
available.

Just because an element is currently widely used for a purpose we deem
inappropriate doesn't mean that its appropriate uses aren't important.
Tables are widely used for layout; <br>-s are widely misused.  Both of
those clearly have other valid uses, so are still in HTML.

> Just because HTML5 redefines the element does not mean that the
> element will suddenly be semantic. Even if people start using it
> purely semantically from now on (and what is the chance of that?), the
> existing websites still carry small-tags that are not compliant with
> the new definition.

Yes.  But the suggested alternative was to deprecate <small> entirely
and invent a new element to convey the semantic of 'small print'.  That
would of course make _all_ current uses of <small> non-conforming.
Presentational <small>-s are going to be non-conforming either way;
allowing semantic <small>-s to conform doesn't change that.

> By redefining it the (existing) web "breaks"; allbeit purely in the
> semantic area. 

That's intentional.  If anybody checks legacy content against the new
standard they will discover that what they did is no longer recommended.
However, browsers will 100% support it and continue to render it as it
always has been, so the 'breakage' is no way visible; if the author
chooses not to care about it then no harm is done.

Smylers

PS: Pentasis, please could you send mails that do at least one of
attributing who you're quoting or include In-Reply-To: headers so that
they continue the existing thread rather than starting a new one.
Without either it's rather tedious to have to look up who said the text
you quote.  Thanks.



More information about the whatwg mailing list