Re: Web Fonts

On Aug 18, 2006, at 09:44, David Woolley wrote:

> but its main use, so far, seems to be for misrepresenting character
> codes in order to represent less commercially attractive (to OS and
> browser developers) language scripts, such as many Indian languages,
> on 8 bit character platforms, and legacy sites.

Very true.

Additionally someone could use downloadable fonts with deliberately  
wrong character-to-glyph mappings as a copy-paste poisoning scheme.  
(I am told there are products for deliberately breaking PDF glyph-to- 
character mappings.)

> For body text at modern designer sizes, almost any font will  
> produce the
> same display bitmap, typically because a fallback bit map font is  
> used.

Not anymore. Nowadays the typical case is anti-aliased rasterization  
from the outline.

> For display text, the designer is usually not content with simple font
> rendering,

Really?

> The de facto need to impose DRM on fonts also requires a
> technical skill for designers that is not natural to them.

No DRM. You are assuming that a Web Fonts scheme has to cater to very  
proprietary commercial fonts. It doesn't and the failure to realize  
this has made previous Web Font attempts fail.

Håkon Lie pointed out in his XTech presentation that according to  
research by an unnamed corporation, there are 500 free ($0) fonts on  
the Web. You don't need a DRM scheme with free-as-in-beer fonts. If  
the commercial foundries don't like DRMlessness, their fonts aren't  
used. It turns out that TrueType does have simple restrictions  
management bits for signaling the embedding policy of the font.  
Browsers could respect those, so that people wouldn't accidentally  
use fonts that aren't OK for distribution.

There is, however, a technical problem. TrueType fonts are designed  
to be obtained from trusted sources and contain hinting code that is  
run inside the TrueType engine. It is said that it is possible to  
construct fonts that put the TrueType engine in an infinite loop. (I  
don't know how this works out with embedded Type 42 fonts in PDFs. I  
haven't heard of any real exploits.)

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

Received on Friday, 18 August 2006 09:28:40 UTC