W3C home > Mailing lists > Public > www-style@w3.org > August 2004

Re: [css4-text] 'text-autospace' and French guillemets

From: Etan Wexler <ewexler@stickdog.com>
Date: Mon, 9 Aug 2004 00:39:28 -0700
To: www-style@w3.org
Message-Id: <3EC22671-E9D7-11D8-A551-000502CB1B77@stickdog.com>

Alexander Savenkov wrote to <mailto:www-style@w3.org> on 8 August 2004 
in "Re[2]: [css4-text] 'text-autospace' and French guillemets" 
(<mid:1236105387.20040808140352@xmlhack.ru>):

>> Do you mean using characters to produce certain glyphs? I canít answer
>> the question because I don't understand it.
>
> That was a rhetorical question, Etan.

Regardless, I don't understand what you meant by "typesetting". I 
understood that you *didn't* mean the physical placement of metal 
slugs. Beyond that, "typesetting" could refer to the mere use of coded 
characters or to the entirety of languages like CSS and Tex.

> If you have the following ď3 %Ē in your text (minus the quotes), the
> right to way to handle the sequence is not to mark it up, but to place
> the special typographical characters that will do the job:
>
> 3,EN SPACE,% or more precisely 3,ZWNBSP,EN SPACE,ZWNBSP,%.

I won't object forcefully. But I wish to make a few notes:

The character U+FEFF, zero width no-break space (ZWNBSP), is to be used 
only as a byte-order mark. Its other semantic as a formatting character 
is deprecated in favor of the character U+2060, word joiner (WJ). So 
the appropriate document fragment would be as follows.

     3&#x2060; &#x2002; &#x2060;%

Even if the editing and storage environments supported the direct use 
of the characters (instead of references to characters), it would be 
cumbersome. Consider the following simple declaration, which could be 
written once and still serve any number of documents.

     text-autospace: numeric-percent alpha-percent;

Now consider the alternative: entering or generating a sequence of 
obscure characters for every percentage sign. The declaration would 
work with legacy content but not with legacy user agents. The character 
entry would require modification to legacy content and does not work 
with most fonts.

When the fonts and/or layout engine don't support the necessary 
characters, the rendered result will be a series of unintelligible 
boxes. I find it worse to see a series of boxes than I find it to see a 
lack of space or a space of the wrong width.

> You donít need to do this:
>
> 3<span class="percentsign">%</span>
> and apply .percentsign { margin-left: .5en; }.
>
> See, itís not the level of markup, itís still the level of plain text.

If that was an argument, I have yet again failed to understand. 
Certainly, when one avoids tagging and instead uses data characters or 
character references, the issue is at the level of plain text.

> Exactly, what space is put between the text and the quote marks in
> France?

I don't know. Some person at the French Academy probably knows. 
(Contact information is at 
<http://www.academie-francaise.fr/contact/index.html>.)

It could be argued that none of markup, style sheets, or textual 
content are the solution. The argument would hold that it is the 
responsibility of the font designer to provide appropriate kerning for 
various languages. What do people think of that?

-- 
Etan Wexler.
Received on Monday, 9 August 2004 07:41:46 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 27 April 2009 13:54:32 GMT