W3C home > Mailing lists > Public > www-style@w3.org > November 2003

Re: CSS21 @font-face removal

From: Henri Sivonen <hsivonen@iki.fi>
Date: Thu, 6 Nov 2003 22:14:25 +0200
Cc: www style <www-style@w3.org>, W3c I18n Group <w3c-i18n-ig@w3.org>
To: Tex Texin <tex@i18nguy.com>
Message-Id: <D13A308E-1095-11D8-B2EB-003065B8CF0E@iki.fi>

On Tuesday, Oct 21, 2003, at 08:21 Europe/Helsinki, Tex Texin wrote:

> I am afraid I don't understand why most everyone is so gloomy on the 
> prospect
> of a font being sufficient to enable users of minority scripts.

I'm gloomy about font embedding/automatic downloading, because:
  * Considering the precedent with non-CJK Asian scripts, I think font 
embedding
    is likely to be abused in a way that actually harms Web accessibility
    of some scripts. (Prolonging the Latin gibberish approach.)

  * Internet Explorer has a large installed base and it has working 
support
    for font embedding at least for the Windows-1252 repertoire. That is,
    there's already a large population of users with a user agent that 
supports
    font downloading. Still, as far as I know, font embedding isn't used 
much
    by Western designers. It's not like designers shy away from features 
the
    just because they are IE-specific. Perhaps there is something 
unattractive
    about font embedding that outweighs the attractiveness of visual 
control.
    Something that runs deeper than the currently available method being
    IE-specific.

  * Font foundries don't want their fonts distributed for free for 
everyone.
    Compared to some other jurisdictions which don't have any official 
copyright
    relaxation for fonts, the U.S. Copyright Office stance that a 
rendering
    produced with a font is not subject to copyright is actually rather 
lax.
    However, because of this, the font foundries tend to take the stance 
that
    hinted fonts constitute software, which in turn has exceptionally 
strong
    corporate protection. Even if font embedding was technically 
feasible, what's
    the point in putting effort into implementing something that almost 
no one
    dares use because of legal issues?

  * Designers tend to pick fonts they believe "everyone" has. This 
includes
    mainly Times New Roman, Arial, Verdana and sometimes Georgia. Western
    designers in general aren't even using the full font repertoire of 
Windows
    and Mac OS, because they seem to think consistency across "all" 
environments
    is more important than picking something less trite that might look
    different for different readers. Since designers wouldn't be able to 
make
    sure embedded fonts work *everywhere*, why would the designers 
choose to
    embed different fonts if they don't dare to use fonts that a 
significant
    number of people (but not the supposed "everyone") already has 
installed?
    What's the point in putting effort into font downloading until
    designers embrace the idea that it is OK to have different 
fontifications
    in different environments? (I think it is unrealistic to expect that
    CSS WG-endorsed font referencing would guarantee the same 
fontification
    everywhere and alleviate the designer concern for sameness.)

  * Fonts are rather large (in bytes) relative to the size of usual Web 
pages.
    The time it takes to download the font may be perceived to be too 
long
    compared to downloading the text. I don't believe font embedding 
would
    actually in practice solve anything for people who want to read 
Chinese in a
    university in England. The CJK font sizes are just too large compared
    to the size of the main content of usual web pages. Also, does 
anyone want
    a flash of unfontified content (compare with the flash of unstyled 
content)?
    Deferring the display of the page until the font has arrived would 
be a bad
    solution. However, a sudden font change after the reader has started 
reading
    the page would be bad, too.

-- 
Henri Sivonen
hsivonen@iki.fi
http://www.iki.fi/hsivonen/
Received on Thursday, 6 November 2003 15:14:28 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 27 April 2009 13:54:25 GMT