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

My 2 cents worth (although no longer legal tender here ;) )

For me the whole issue of <b> and <i> being semantic elements tends to
confuse the distinction between semantic and non-semantic elements.

The status of <b> and <i> being semantic seems to based on the need of
being semantic meaning distinct from <em> and <strong> to a series of
usages limited to specific typesetting traditions in Latin and
Cyrillic (and maybe by extension modern Greek).

Both <i> and <b> can be used semantically in multiple ways, and
nothing in the tag by itself indicates wether the author is using the
tag semantically or presentationally, and if the author is using the
tag semantically, there is nothing in the tag to indicate the distinct
semantic meaning the author is applying to the tag (other than it is
semantic ;)  )

For the languages I currently work on, how would I style <i> and <b>
elements if i were suing those elements in a semantic way?

* Never use an italic face, there is no italic face for the scripts I work with
* Maybe use a bold typeface, but many of the fonts I use and are
readily available do not have a bold weight, they come in a single
weight. So bold typefaces are synthetic and don't really work well
* For the languages I work with underlining text is a traditional way
of applying semantic meaning attached to <em>, <i>, <strong> and <b>
elements as spec currently stands
* Change in typeface
* Change in colour of text
* Highlighting text

But I suspect that making <u> semantic will just confuse the issue any
more, since I might have to style <em>, <i>, <strong>, <b> and <u> all
the same way.

On 27 January 2011 11:33, Kang-Hao (Kenny) Lu <kennyluck@w3.org> wrote:
> Hi HTML WG,
>
> Sorry for bring up this old and non-technical issue.
>
>>  This should be separated out into two separate proposals.
>
> I decided on one of the two suggestions I proposed, that is, make all
> <b>,<i>  and<u>  non-semantic.[1] That is,<b>  =<i>  =<u>  =<span>
> semantically. I hope this won't be considered non-orthogonal to the
> raised issue as this affects<b>  and<i>  as well.
>
> The majority of people just sense certain inconsistency about the
> current status of<b>,<i>  and<u>, and there are certainly lots of
> other similar proposals, but I might not be motivated to write up any of
> these:
> - Redefine<u>  as a semantic element.
> - Reintroduce<u>  as a presentational element or element with no
> semantics in HTML, potentially having the status "obsolete but
> conforming". Do nothing to<b>  and<i>.
> - If<u>  <b>  <i>  have the class attribute on, they are semantic.
> Otherwise, they are presentational or elements with no semantics.
> - bra, bra
>
> The editor claimed that<b>  and<i>  have solid and common use cases[2].
> I think this might be a valid point as long as
> 1. The spec text "..., or some other prose whose typical typographic
> presentation is italicized."[3] is removed from the spec. I can't
> imagine how to translate this part of the spec into a language that
> doesn't use italic type often, say, Chinese or Japanese. Italic type is
> a non-trivial typography effect.
> 2. There are enough real world visual examples that actually style<b>
> and<i>  differently. I would be curious to see how they are styled. Any
> pointer?
>
> Any comments?
>
> [1] http://www.w3.org/html/wg/wiki/ChangeProposals/BIUArePresentational
> [2] http://www.w3.org/Bugs/Public/show_bug.cgi?id=10838#c3
> [3] http://www.w3.org/TR/html5/text-level-semantics.html#the-i-element
>
> Cheers,
> Kenny
>
>
>
>
>



-- 
Andrew Cunningham
Senior Project Manager, Research and Development
Vicnet
State Library of Victoria
Australia

andrewc@vicnet.net.au
lang.support@gmail.com

Received on Thursday, 27 January 2011 01:31:50 UTC