W3C home > Mailing lists > Public > www-style@w3.org > May 1998

Re: Embedding fonts

From: Chris Lilley <chris@w3.org>
Date: Tue, 12 May 1998 15:36:44 +0200
Message-ID: <3558506C.267612A8@w3.org>
To: Chris Wilson <cwilso@MICROSOFT.com>
CC: "'Todd Fahrner'" <todd@lowbrow.com>, www-style@w3.org, www-font@w3.org
Chris Wilson wrote:
> 
> BTW, IE's downloadable font implementation is a strict subset of what CSS2
> defines for Fonts.  We didn't pick that syntax randomly - it came straight
> from the WebFonts Working Draft (http://www.w3.org/TR/WD-font), which both
> Microsoft and Bitstream had a hand in developing.  (Todd, I know that you
> know this; this was for others' benefit.)

Yes, thanks for pointing that out Chris.

> > > Todd Fahrner [SMTP:todd@lowbrow.com] wrote:

> >I haven't heard anything to suggest that Netscape will be supporting
> >OpenType anytime soon. 

s/Opentype/Embedded Opentype/g

 Nor I. They do have a plug-in font rendering architecture though, so a
third party could implement it.

> However, TrueDoc is now usable in Internet Explorer
> >through an ActiveX control.[1] How do you and your readers like ActiveX?
> >You have to embed the control in the HTML of every page you want to use
> >TrueDoc in - can't specify it in the stylesheet and have done.

I think you just answered your own question there ;-)

> This is indeed true, and is not an optimal solution.  

> Todd, you fail to mention two of the main drawbacks of the TrueDoc format -
> 1) it throws away all hinting and auto-hints the font - some people say they
> can see the difference, and 2) the last time I looked, the TrueDoc recorder
> doesn't pay any attention to the font license information, e.g. the bit in
> the font that says, "you're not allowed to embed this." 

Presumably because they do not see themselves as embedding the font, but
as embedding the output of the font. (Again, that was for other folks
benefint, I know that Chris and Todd are aware of the argument).

> Our OpenType
> embedding does; if the typographer has a restrictive license, you will not
> be able to embed it.  Designers need to realize that fonts are software (as
> well as works of art), and are protected as such.

Yup. Embedded fonts need to be created with regard to their license
(some foundries let you, some don't); transmitted in a form that is not
directly usable, and not installed or otherwise made available to people
to use for other documents.

> >Unfortunately, Netscape (the TrueDoc implementor) doesn't implement enough
> >CSS to let you embed fonts through CSS. You still have to muck around in
> >the HTML, against the recommendations of the HTML 4.0 Specification.
> 
> Is this really against the recommendation of the HTML4 spec?  It's a bit of
> a hack, to be sure, but they just use a META tag, which I don't think is
> very strenuously defined in HTML 4.0.

Their syntactic usage of the META tag is indeed within the spec,
although their usage of tagging for font embedding is contrary to the
Web Fonts design principles which were decided by a number of companies,
Netscape and Microsoft included. Their syntactic usage of the deprecated
face attribute on font is also within the spec for Transitional HTML
4.0. Of course, that attribute was added because certain vendors had
already implemented and it was in widespread use; the presence of the
face attribute in Transitional HTML 4.0 representss a documentation of
existing practice not a recommendation for creation of new content.

> >I hope this will be fixed in an upcoming 5.0/Mozilla release. It's a little
> >late, but I'd also like to see some way to specify a list of fonts in
> >different formats in CSS @font-face rules,

You mean, apart from the existing way to specify a list of fonts in
different formats in CSS @font-face rules ? ;-)

>  for UAs to sort out, so CSS
> >authors aren't forced to take sides in a proprietary font format war.
> >Today, if you use OpenType, you cast your vote for Microsoft and their
> >errant (IMO) type rendering. If you use TrueDoc, you can be browser-neutral
> >and pro-anti-aliasing,but only by means of ActiveX and deprecated HTML in
> >place of CSS.
> 
> Hey, someone should just step up and volunteer to implement the correct CSS2
> syntax in Mozilla. 

;-) I have spoken to the Netscape folks about that, they do promise to
change it in the next version. 

It would be good to see fuller support for the correct CSS2 WebFont
syntax in new versions of all the CSS-supporting browsers, actually. I
am thinking particularly of 

a) the intelligent font matching (based on the Panose number)
b) the font-size-adjust
c) the font-stretch property for specifying condensed and expanded faces
in the same family.

>  I could probably convince our typography team to assist
> anyone who wanted to implement OpenType embedding in Mozilla as well.  I'm
> sure you can imagine why I can't offer to do it myself.  ;^)

That assistance would include help with the necessary licensing of the
Microtype Express technology from Agfa?

> You can, in fact, specify a list of different formats in the CSS @font-face
> rules for the UA to select. 

Yes, exactly (tsk, Todd, you know that already) although Navigator 4.0x
will not recognise the syntax.

> Even so, though, the designer would need to
> create two copies of the font (one as Embedded OpenType, one as TrueDoc),

right, but you can point to these two copies from your style sheet and
say which format each uses. I agree having to make two or more formats
is a pain.

> presuming you didn't want to insert the TrueDoc ActiveX control in your
> pages.  If you're going to build two copies of the font anyway, you can use
> the CSS2 syntax IE4 uses for the EOT file, and the META syntax Nav uses -
> neither will interfere with each other.

Actually they will. Not the meta syntax per se, which is harmless
(although the hxburned attribute inserted by Hexmac Typograph is
non-standard) but the fact that you need to have font face tags all over
your HTML, which has several problems:

- tying the content down to a single style of presentation
- high specificity of stylistic control on a single element, messes up
inheritance in the CSS
- people omit the actual 'semantic' tags and just use font, making the
pages inaccessible, hard to search, and difficult to automatically
convert to a style sheet
- encourages using fonts to lie about content/charset encodings; HTML
and CSS take rather different approaches toward the byte/character/glyph
model.

But, it is possible to use the CSS2 syntax for clients that want eot,
pfr, or any other flavour of font; in the same style sheet, without
ambiguity. WebFonts was explicitly designed that way, because we knew
there were multiple font formats now and would likely be more in the
future.

-- 
Chris Lilley, W3C                             http://www.w3.org/
Graphics,Fonts,Stylesheets Guy    The World Wide Web Consortium
http://www.w3.org/people/chris/              INRIA,  Projet W3C
chris@w3.org                       2004 Rt des Lucioles / BP 93
+33 (0)492 387 987 <NEW    06902 Sophia Antipolis Cedex, France
Received on Tuesday, 12 May 1998 09:37:27 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 27 April 2009 13:53:55 GMT