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

Re: CSS21 @font-face removal

From: David Woolley <david@djwhome.demon.co.uk>
Date: Tue, 21 Oct 2003 07:44:08 +0100 (BST)
Message-Id: <200310210644.h9L6i8401737@djwhome.demon.co.uk>
To: www-style@w3.org

> Of course, there are only two implementations of CSS in the world ....
> and yes, I suspect that "HTML+CSS browsers we are familiar with" was
> the reason it was dropped. Which is a poor reason, unless CSS 2.1

In terms of the interoperability rule being a test of lack of ambiguity
and general implemtability, it may be flawed, but it does make sense in
terms of encouraging real world use of CSS standards, rather than the
use of the many commercial publications that purport to say what real
world CSS is.

> Yes, it does. Furthermore, it impacts mobile web users more than
> others because the 'everyone has fonts by these names' assumption
> falls down flat there.

My impression is that general web page designers aren't really aware of
font downloading.  Part of this is chicken and egg; the popular text books
don't mention it because there is no interest and there is no interest
because most designers only read popular text books, etc., and never the
source documents.  However, I think it is also because text as GIF gives
them more control and they want to do more graphical manipulation on text
("Word Art") than CSS allows.  (The last point could change if and when
the dominant web browser ever implements SVG natively [3].)

> DW> I don't think there is much awareness of the feature even in such
> DW> communities.
> 
> Again, you would need to demonstrate that.

This would need a major survey.   I certainly don't have the personal
funds for that.  One is trying to prove a partial negative here.

An indication of general lack of awareness of font embedding is the 
number of PDF documents that I've seen that fail to embed non-standard
fonts that they use, even though font embedding is a strong feature
of PDF.

> 
> DW>  The only example I've seen was a Symbol font hack

Note others refer to "Symbol font hack" as "Latin1 gibberish" in this
thread.

> DW> (misrepresenting ISO 8859/1 characters as glyphs for something else)
> DW> for Telugu.
> 
> I suggest you look harder if you have seen only a single example.

This was a few years ago, but I was trying to find an example (not
specifically foreign languages) of the mechanism being used in anger,
rather than as part of a demonstrator page, to show to colleagues as
something to consider.  There may be a few more real life examples now,
but I've failed to see them used anywhere in anger, even in places where
they are of obvious benefit.  I have used them myself, though, although
on a page only published to a language class that I attend.

> Since when is Chinese a minority language? Last I looked it was the
> world's number one language.

Chinese is very much a minority language in the context which I quote
which is for users, including people with Cantonese as their mother
tongue[1], in Western coutries, as far as the likelihood of having
a web browser with appropriate fonts installed.  To that extent, 
some of the sites providing learning resources on Chinese language
and culture resort to imaging the text, at least for some of their
pages, and most other feel the need to explain how to get fonts
(often with questionable advise, like using Chinese language bolt ons,
and using the font from Microsoft that is only actually licensed to
people who have Microsoft Office (no-one reads click through licences).

None of these sites use font downloading.  The BBC are also an example.
In the case of Chinese they do assume an audience with fonts installed,
but for Hindi (Indian professionals tend to use English) they have a GIFed
message advising people to download a font, which leads to a licence
agreement.  Hindi, in the target area, is not a minority language in a
normal sense, but Hindi font support tends to require the very latest
software (the script is nearly all ligatures, but is not encoded as
glyphs, so the font engine has to support this - the X implementation
of Hindi under Mozilla doesn't implement the ligatures, at least not a
couple of years ago).  The BBC do properly character set code their pages.

[1] I'm learning Chinese and most of the class either have Cantonese as
their mother tongue or come from such a family, but very few of them
have fonts installed.  One, who actively uses them, uses 16x16 fonts from
NJSTAR, not the Windows[2] support.  One had them installed by me, when
I was actually installing something else (not someone from a Chinese
background).  One may have actually used a system with them installed or
auto-upgraded[4], but the majority seem not to have any installed.  This is
even more the case when they are using office machines.

[2] we're talking non-technical users here, and a country where every
PC sold has Windows, so it is a Windows only environment.

[3] Native support is even more important post EOLAS, although 
one might just get away with data: URLs.

[4] Chinese is a free download for IE4+ prior to Windows 2000 and is
on the distribution media, but not a default install post Windows 2000,
in Western versions of Windows.  Having it on the CD is probably worse,
as various mass market PC suppliers in the UK charge extra for installation
media and IT departments tend to hold it in companies, even if users are
allowed to use Windows Upgrade.
Received on Tuesday, 21 October 2003 02:44:24 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 27 April 2009 13:54:24 GMT