Re: ISSUE-144 (conforming-u): Chairs Solicit Proposals

On Wed, 26 Jan 2011, Henri Sivonen wrote:
> 
> Kang-Hao (Kenny) Lu had already drafted a CP, so I amended his CP 
> instead.
> 
> (The draft is at 
> http://www.w3.org/html/wg/wiki/ChangeProposals/UShouldBeConforming ; 
> I'll leave it to Kenny to submit it officially.)

This CP mentions two use cases.

The first is as the proper name mark in Chinese. Based on my research, it 
seems this is actually quite rare, and equivalent more to Western 
typographic conventions such as overlines on recurring decimals. As such, 
it seems like a use case already handled by U+0332, much like overlines 
are handled by U+0305.

The second is for indicating spelling mistakes. This is an interesting 
idea (I haven't seen it suggested before). It seems somewhat at odds with 
legacy use of <u> (significantly more so than <small>'s new definition 
relative to legacy use of <small>, for instance). It would be helpful to 
know how common it is for Web pages to use an element to indicate spelling 
problems. It doesn't seem to be something I've seen much, anecdotally.

Also, the CP implies that the definitions of <b>, <i>, <s>, and <small> 
in HTML now are in bad faith -- that they are definitions intended to 
cover an embarassment; maybe an excuse for allowing elements under a 
pretext different than the actual rationale. I must emphatically point 
out that this is simply not the case. The current definitions are not "fig 
leaves"; they are intended to address real problems that are commonly 
seen on Web pages, in a pragmatic manner that is backwards compatible with 
legacy user agents and to a significant extent consistent with legacy 
content. It's possible that the definitions need some tightening up, as 
the CP suggests, but to that end I would recommend that people file bugs 
on the offending ambiguities.


I urge the proponents of this CP to consider why their arguments do not 
apply to <font>, <big>, <layer>, <blink>, <tt>, <center>, align="", etc, 
or if they do, to be consistent in their proposal and reintroduce all 
these elements. Even worse than turning HTML back into the presentational 
language of the 90s would be to have HTML turn into a design-by-committee 
mess of inconsistent decisions based purely on what elements we bothered 
to consider in the escalation process.

-- 
Ian Hickson               U+1047E                )\._.,--....,'``.    fL
http://ln.hixie.ch/       U+263A                /,   _.. \   _\  ;`._ ,.
Things that are impossible just take longer.   `._.-(,_..'--(,_..'`-.;.'

Received on Wednesday, 26 January 2011 21:35:28 UTC