Re: CSS21 @font-face removal

On Tuesday, Oct 21, 2003, at 08:55 Europe/Helsinki, L. David Baron 
wrote:

> On Monday 2003-10-20 19:52 +0300, Henri Sivonen wrote:
>> Then there's the practice of transferring Latin gibberish and applying
>> a font that is a Latin font from the system's point of view but
>> contains glyphs for another script.
>
> Although it's getting a little off-topic, it's worth noting that about
> half the bugs filed to request that Mozilla support dynamic fonts were
> filed because of such sites.

Which is why I was so quick to assume the Latin gibberish approach when 
I saw the suggestion that font downloading helped with minority 
scripts. It's also why the answer to "Does Mozilla support downloadable 
fonts?" in the Mozilla Web Author FAQ[1] is not just "Downloadable 
fonts are not supported." but points out that the author should 
probably want Unicode instead.

>> I think CSS 2.1 should not
>> accommodate fontifying Latin gibberish to look like text in a minority
>> script in browsers that happen to support such a trick. That approach
>> may appear to work (for some value of "work") in some cases but causes
>> problems with search engines and usually with browsers other than the
>> one the author of the page was using.
>
> I'm not sure how a browser could prevent it -- how could it detect that
> a font doesn't contain the characters it claims to contain?  (Or did 
> you
> mean something else by "CSS 2.1 should not accomodate..."?)

I meant that enabling the Latin gibberish trick should not be used as a 
rationale for including a feature in CSS 2.1. (In other messages in 
this thread it was pointed out that it wasn't offered as a rationale.)

>  And should
> charset=x-user-defined just be rejected, or should it be left for 
> people
> who want to exchange documents containing characters that really aren't
> in Unicode?

Does x-user-defined have a clear meaning? Is there a well-defined way 
to process it? Does it just mean that the author wants the browser to 
pass the bytes through to a byte-equals-character Windows API as if the 
encoding was Windows-1252 and then wants to redefine the glyphs for the 
characters?

Using characters that aren't in Unicode isn't a good idea on the Web. 
If the commucating parties have agreed on the meaning of the 
non-Unicode characters (unlikely to happen on the Web), they can use 
the PUA for private communication.

[1] http://www.mozilla.org/docs/web-developer/faq.html#downloadablefonts

-- 
Henri Sivonen
hsivonen@iki.fi
http://www.iki.fi/hsivonen/

Received on Thursday, 6 November 2003 14:15:48 UTC