New work on fonts at W3C

Previous Topic Next Topic
 
classic Classic list List threaded Threaded
272 messages Options
1234 ... 14
Reply | Threaded
Open this post in threaded view
|

New work on fonts at W3C

Chris Lilley
Hello www-style,


Introduction

W3C is collecting ideas for work related to downloadable fonts on the Web. This email summarizes the current situation, and asks for feedback on a draft charter [0] for a future W3C Font working group or interest group. Please send feedback on the charter to the publicly archived mailing list [hidden email].

Background

The method of linking to fonts from CSS stylesheets or from SVG content - WebFonts - is already standardized in CSS2 [1] and SVG 1.1 [2]. That work is being refined in the CSS3 Fonts module [3], for the CSS serialization. SVG 1.1 allows both CSS and XML serialisations of WebFonts. SVG Tiny 1.2 uses the XML serialisation. XSL may use the WebFont mechanism in future, in the XL serialization.

CSS2 WebFonts allows for font matching, font synthesis, and font download. of those, only font download has been implemented in multiple user agents.

What has hindered deployment, however, has been the choice of what font format to download.

CSS2 suggests several formats, none of them being required for compliance. Microsoft implemented WebFonts in IE4, using the then-proprietary EOT encoding of an OpenType font. EOT (and the related MicroType Express format) have since been documented in a submission to W3C [4] with the promise of royalty-free commitments, if the EOT and MicroType Express formats form part of a W3C Recommendation.

SVG requires one format (SVG fonts) and allows others to be linked. Adobe implemented WebFonts in ASV, supporting SVG and CFF (a subset of OpenType, using only Type 1 glyphs). Apache implemented WebFonts in Batik, supporting the SVG format. Opera implements WebFonts with SVG, and the fonts may be applied to SVG, HTML, or mixed SVG/XHTML content.

Several Web browsers now implement WebFonts, with the OpenType format: Firefox 3.5 (beta 4 was tested), Safari 3.1 and 4, and Opera 10 (alpha). W3C has produced tests for complex international script rendering with downloadable fonts; the test and results are available [5]. (Firefox 3.5 beta 4 results are yet to be added to this table).

Font licenses vary. Some fonts, libre, are provided with source code and permit extension, modification, and redistribution. The Open Font Library [6] collects such libre fonts. Other fonts may be downloaded and used, but may not be changed, subsetted, or converted to another format. (Often such fonts are contained in zip files; being able to link to a specific font inside a zipfile is desirable).

Some fonts are licensed to a specific site or domain. EOT provides one way to indicate this in the font itself. Cross Origin Resource Sharing (CORS, previously known as Access Control) is a W3C specification which may also be used to indicate this [7]. Mozilla Firefox restricts downloadable OpenType fonts to those permitted by CORS.
There may be other ways to indicate metadata, so that foundries and font licensees may indicate the nature of their agreement.

Lastly there are of course fonts intended for use exclusively with print, whose licenses explicitly preclude use on the Web.

Please take a moment to read and comment on the proposed charter [0] which is a work in progress.

[0] http://www.w3.org/2009/03/fonts-wg-charter
[1] http://www.w3.org/TR/2008/REC-CSS2-20080411/fonts.html
[2] http://www.w3.org/TR/SVG11/fonts.html
[3] http://dev.w3.org/csswg/css3-fonts/
[4] http://www.w3.org/Submission/2008/01/
[5] http://www.w3.org/International/tests/tests-html-css/tests-webfonts/results-font-linking
[6] http://www.openfontlibrary.org/
[7] http://www.w3.org/TR/access-control/

--
 Chris Lilley                    mailto:[hidden email]
 Technical Director, Interaction Domain
 W3C Graphics Activity Lead
 Co-Chair, W3C Hypertext CG


Reply | Threaded
Open this post in threaded view
|

Re: New work on fonts at W3C

Jonathan Kew-2
Regarding http://www.w3.org/2009/03/fonts-wg-charter,

there is a small edit needed in the middle of section 1, under "The  
goals for this font format are:"

• to allow as {broad/wide} a range of implementations (including "open  
source") as possible, ....
       missing: ^^^^^^^^^^

JK


Reply | Threaded
Open this post in threaded view
|

Re: New work on fonts at W3C

John Daggett
In reply to this post by Chris Lilley
> W3C is collecting ideas for work related to downloadable fonts on the
> Web. This email summarizes the current situation, and asks for feedback
> on a draft charter [0] for a future W3C Font working group or interest
> group. Please send feedback on the charter to the publicly archived
> mailing list [hidden email].

What is Microsoft's view on the proposed Fonts WG charter in it's
current form?

John Daggett
Mozilla Japan

Reply | Threaded
Open this post in threaded view
|

Re: New work on fonts at W3C

germain (Bugzilla)
In reply to this post by Chris Lilley
Le mardi 12 mai 2009, Chris Lilley a écrit :
> Several Web browsers now implement WebFonts, with the OpenType format:
> Firefox 3.5 (beta 4 was tested), Safari 3.1 and 4, and Opera 10 (alpha).

Hello,
you may wish to add Konqueror 4.3 to this list, with preliminary support for
Web Fonts being available in the KHTML engine in beta 1 (released today), and
better support coming in the next beta.

At the moment, We support downloadable fonts in Open Type and True Type
format - with non-blocking rendering.
Fonts may be compressed (beta 2) using any of gzip, bzip or lzma compression.

Greetings,
Germain

Reply | Threaded
Open this post in threaded view
|

RE: New work on fonts at W3C

Sylvain Galineau
In reply to this post by John Daggett
> What is Microsoft's view on the proposed Fonts WG charter in it's
> current form?
I know of no objections at this  time. On the editorial side, I'd suggest adding link to the EOT submission request [1] for completeness since the technology it describes is both cited and highly relevant. It's totally fine to make it clear this is just an informational reference and does not constitute a starting point for the WG, of course.

[1] http://www.w3.org/Submission/2008/01/
Reply | Threaded
Open this post in threaded view
|

Re: New work on fonts at W3C

Anne van Kesteren-2
In reply to this post by Chris Lilley
On Tue, 12 May 2009 16:07:59 +0200, Chris Lilley <[hidden email]> wrote:
> Some fonts are licensed to a specific site or domain. EOT provides one  
> way to indicate this in the font itself. Cross Origin Resource Sharing  
> (CORS, previously known as Access Control) is a W3C specification which  
> may also be used to indicate this [7]. Mozilla Firefox restricts  
> downloadable OpenType fonts to those permitted by CORS.
> There may be other ways to indicate metadata, so that foundries and font  
> licensees may indicate the nature of their agreement.
>
> [7] http://www.w3.org/TR/access-control/

Just to be clear: CORS is not about license enforcement. It is about  
alleviating the same-origin policy in certain scenarios. (Whether the  
same-origin policy should apply for fonts at all is something I'm not sure  
about.)


--
Anne van Kesteren
http://annevankesteren.nl/

Reply | Threaded
Open this post in threaded view
|

Re: New work on fonts at W3C

Chris Lilley
On Monday, May 18, 2009, 10:09:13 PM, Anne wrote:

AvK> On Tue, 12 May 2009 16:07:59 +0200, Chris Lilley <[hidden email]> wrote:
>> Some fonts are licensed to a specific site or domain. EOT provides one  
>> way to indicate this in the font itself. Cross Origin Resource Sharing  
>> (CORS, previously known as Access Control) is a W3C specification which  
>> may also be used to indicate this [7]. Mozilla Firefox restricts  
>> downloadable OpenType fonts to those permitted by CORS.
>> There may be other ways to indicate metadata, so that foundries and font  
>> licensees may indicate the nature of their agreement.

>> [7] http://www.w3.org/TR/access-control/

AvK> Just to be clear: CORS is not about license enforcement.

Right, its about allowing resources from one domain to be used from another, if they would otherwise be disallowed due to security policies.

AvK> It is about  
AvK> alleviating the same-origin policy in certain scenarios. (Whether the
AvK> same-origin policy should apply for fonts at all is something I'm not sure
AvK> about.)

I didn't express an opinion on whether it *should be* only that at least one implementation *does* currently impose a same-origin policy for webfonts, and does allow CORS to be used to widen that policy.



--
 Chris Lilley                    mailto:[hidden email]
 Technical Director, Interaction Domain
 W3C Graphics Activity Lead
 Co-Chair, W3C Hypertext CG


Reply | Threaded
Open this post in threaded view
|

Re: New work on fonts at W3C

Anne van Kesteren-2
On Mon, 25 May 2009 18:50:45 +0200, Chris Lilley <[hidden email]> wrote:
> On Monday, May 18, 2009, 10:09:13 PM, Anne wrote:
>> It is about alleviating the same-origin policy in certain scenarios.  
>> (Whether the same-origin policy should apply for fonts at all is  
>> something I'm not sure about.)
>
> I didn't express an opinion on whether it *should be* only that at least  
> one implementation *does* currently impose a same-origin policy for  
> webfonts, and does allow CORS to be used to widen that policy.

Just to be clear, I didn't mean to imply you were. I was just stating my  
own reservation.


--
Anne van Kesteren
http://annevankesteren.nl/

Reply | Threaded
Open this post in threaded view
|

Re: New work on fonts at W3C

Chris Lilley
On Monday, May 25, 2009, 8:01:34 PM, Anne wrote:

AvK> On Mon, 25 May 2009 18:50:45 +0200, Chris Lilley <[hidden email]> wrote:
>> On Monday, May 18, 2009, 10:09:13 PM, Anne wrote:
>>> It is about alleviating the same-origin policy in certain scenarios.  
>>> (Whether the same-origin policy should apply for fonts at all is  
>>> something I'm not sure about.)

>> I didn't express an opinion on whether it *should be* only that at least  
>> one implementation *does* currently impose a same-origin policy for  
>> webfonts, and does allow CORS to be used to widen that policy.

AvK> Just to be clear, I didn't mean to imply you were. I was just stating my
AvK> own reservation.


OK, so we are on the same page, then.

Actually I have reservations myself; I discovered this when a talk I was doing, intended to be live but then saved to local disk in case of network issues, mysteriously stopped working. Mozilla (3.5b4) decided that local disk was not the same as w3.org so refused to download fonts any more.

I guess w3 should add cors for the webfont download tests, to avoid such problems.

--
 Chris Lilley                    mailto:[hidden email]
 Technical Director, Interaction Domain
 W3C Graphics Activity Lead
 Co-Chair, W3C Hypertext CG


Reply | Threaded
Open this post in threaded view
|

Re: New work on fonts at W3C

davelab6
2009/5/26 Chris Lilley <[hidden email]>:
>
> Actually I have reservations myself; I discovered this when a talk I was doing,
> intended to be live but then saved to local disk in case of network issues,
> mysteriously stopped working. Mozilla (3.5b4) decided that local disk was
> not the same as w3.org so refused to download fonts any more.

I think it is essential that web fonts not break "Save Page As" functionality.

Cheers,
Dave

Reply | Threaded
Open this post in threaded view
|

Re: New work on fonts at W3C

Boris Zbarsky
Dave Crossland wrote:
>> Actually I have reservations myself; I discovered this when a talk I was doing,
>> intended to be live but then saved to local disk in case of network issues,
>> mysteriously stopped working. Mozilla (3.5b4) decided that local disk was
>> not the same as w3.org so refused to download fonts any more.
>
> I think it is essential that web fonts not break "Save Page As" functionality.

Hold on.  Hold on.

Saving just the HTML already generally breaks functionality (e.g.
because the base URI changes) in browsers.

Saving the whole webpage (whether as MHTML or the format Gecko currently
uses) works, as long as the UA actually downloads all the data.

Gecko has a known bug wherein data linked from stylesheets is not
downloaded.  This means not only fonts, but background images, border
images, etc, etc.

So if the saving above was in fact "save page as", then you're just
seeing a Gecko bug.  But I suspect the saving was "copy the file over to
my hard drive and adjust the links or set base URI".  Chris, do you mind
disambiguating which it was?

-Boris


Reply | Threaded
Open this post in threaded view
|

Re: New work on fonts at W3C

Chris Lilley
In reply to this post by davelab6
On Tuesday, May 26, 2009, 3:02:37 PM, Dave wrote:

DC> 2009/5/26 Chris Lilley <[hidden email]>:

>> Actually I have reservations myself; I discovered this when a talk I was doing,
>> intended to be live but then saved to local disk in case of network issues,
>> mysteriously stopped working. Mozilla (3.5b4) decided that local disk was
>> not the same as w3.org so refused to download fonts any more.

DC> I think it is essential that web fonts not break "Save Page As" functionality.

Hi Dave,

Yes, I agree.

In this instance, I see it mainly as a Mozilla bug in 'save web page'; URIs in html are rewritten and resources (images, etc) saved locally, but URIs in CSS (images, fonts, etc) are not rewritten and the referenced resources are not saved.

Can anyone point out an existing bugzilla on this, before I create a new, possibly duplicate one, on that?

--
 Chris Lilley                    mailto:[hidden email]
 Technical Director, Interaction Domain
 W3C Graphics Activity Lead
 Co-Chair, W3C Hypertext CG


Reply | Threaded
Open this post in threaded view
|

Re: New work on fonts at W3C

Chris Lilley
In reply to this post by Boris Zbarsky
On Tuesday, May 26, 2009, 3:41:05 PM, Boris wrote:

BZ> Gecko has a known bug wherein data linked from stylesheets is not
BZ> downloaded.  This means not only fonts, but background images, border
BZ> images, etc, etc.

OK so known bug (link, please? I'd like to add a cc).

BZ> So if the saving above was in fact "save page as", then you're just
BZ> seeing a Gecko bug.  But I suspect the saving was "copy the file over to
BZ> my hard drive and adjust the links or set base URI".  Chris, do you mind
BZ> disambiguating which it was?

It was the latter (I saved a w3.org web page locally, then copied parts of it into a foreignObject in a  local svg page, and the links were absolute but then stopped working) but in fact the former also breaks.


--
 Chris Lilley                    mailto:[hidden email]
 Technical Director, Interaction Domain
 W3C Graphics Activity Lead
 Co-Chair, W3C Hypertext CG


Reply | Threaded
Open this post in threaded view
|

Re: New work on fonts at W3C

Chris Lilley
In reply to this post by germain (Bugzilla)
On Wednesday, May 13, 2009, 6:14:27 PM, Germain wrote:

GG> Le mardi 12 mai 2009, Chris Lilley a écrit :
>> Several Web browsers now implement WebFonts, with the OpenType format:
>> Firefox 3.5 (beta 4 was tested), Safari 3.1 and 4, and Opera 10 (alpha).

GG> Hello,
GG> you may wish to add Konqueror 4.3 to this list, with preliminary support for
GG> Web Fonts being available in the KHTML engine in beta 1 (released today), and
GG> better support coming in the next beta.

Thanks, I will add it.

GG> At the moment, We support downloadable fonts in Open Type and True Type
GG> format - with non-blocking rendering.
GG> Fonts may be compressed (beta 2) using any of gzip, bzip or lzma compression.

Excellent, thanks for the details.



--
 Chris Lilley                    mailto:[hidden email]
 Technical Director, Interaction Domain
 W3C Graphics Activity Lead
 Co-Chair, W3C Hypertext CG


Reply | Threaded
Open this post in threaded view
|

Re: New work on fonts at W3C

Boris Zbarsky
In reply to this post by Chris Lilley
Chris Lilley wrote:
> In this instance, I see it mainly as a Mozilla bug in 'save web page'; URIs in html are rewritten and resources (images, etc) saved locally, but URIs in CSS (images, fonts, etc) are not rewritten and the referenced resources are not saved.
>
> Can anyone point out an existing bugzilla on this, before I create a new, possibly duplicate one, on that?

https://bugzilla.mozilla.org/show_bug.cgi?id=115107

-Boris


Reply | Threaded
Open this post in threaded view
|

Re: New work on fonts at W3C

Chris Lilley
In reply to this post by Jonathan Kew-2
On Wednesday, May 13, 2009, 10:12:25 AM, Jonathan wrote:

JK> Regarding http://www.w3.org/2009/03/fonts-wg-charter,

JK> there is a small edit needed in the middle of section 1, under "The  
JK> goals for this font format are:"

JK> • to allow as {broad/wide} a range of implementations (including "open
JK> source") as possible, ....
JK>        missing: ^^^^^^^^^^

Thanks Jonathan, will fix.





--
 Chris Lilley                    mailto:[hidden email]
 Technical Director, Interaction Domain
 W3C Graphics Activity Lead
 Co-Chair, W3C Hypertext CG


Reply | Threaded
Open this post in threaded view
|

Re: New work on fonts at W3C

Chris Lilley
In reply to this post by Sylvain Galineau
On Friday, May 15, 2009, 2:36:59 AM, Sylvain wrote:

>> What is Microsoft's view on the proposed Fonts WG charter in it's
>> current form?
SG> I know of no objections at this  time. On the editorial side, I'd
SG> suggest adding link to the EOT submission request [1] for
SG> completeness since the technology it describes is both cited and
SG> highly relevant. It's totally fine to make it clear this is just
SG> an informational reference and does not constitute a starting point for the WG, of course.

SG> [1] http://www.w3.org/Submission/2008/01/

Yes, I agree that its cited and relevant and should be linked. I will add it.

I also agree that the clarification that this is not necessarily the start point should be added.


--
 Chris Lilley                    mailto:[hidden email]
 Technical Director, Interaction Domain
 W3C Graphics Activity Lead
 Co-Chair, W3C Hypertext CG


Reply | Threaded
Open this post in threaded view
|

RE: New work on fonts at W3C

Sylvain Galineau
In reply to this post by Chris Lilley
Note that Ascender published a web font proposal last week that relates to this charter: http://blog.fontembedding.com/post/2009/06/10/New-Web-Fonts-Proposal.aspx



Reply | Threaded
Open this post in threaded view
|

Re: New work on fonts at W3C

Mikko Rantalainen
Sylvain Galineau wrote:
> Note that Ascender published a web font proposal last week that
> relates to this charter:
> http://blog.fontembedding.com/post/2009/06/10/New-Web-Fonts-Proposal.aspx

About that proposal...

"However almost all commercial fonts are licensed only for desktop use,
under licenses that do not allow posting to web servers."

There's no technological problem. The only problem is the license for
those commercial fonts! Creating a yet another obfuscated font format
does not change that fact that you still cannot use those fonts because
they are licensed for desktop use only!


"Commercial font developers are unwilling to allow their fonts, licensed
for use on desktops, to be posted on the web."

Again, this is their choice. Why would creating yet another format
change this a bit? They own the font, they decide how it can be used.


"Most font developers believe that without a technological check-point
(even a simple one), that web developers and server owners will not
understand that they may not simply copy a font from a workstation and
use it on the web."

Please, forward this to tech evangelism department. The key words here
are "most font developers believe". If the font developers believe in
flying spaghetti monster, creating another font format will not help
with that, either. The only real choice is to explain the situation in
terms they can understand. There's no and will not be an effective DRM
system!


"a technological check-point (even a simple one), that web developers
and server owners will not understand"

Do we really want to endorse any technology which has the key merit of
being too hard to understand to web developers and server owners?



Also note that browsers from multiple vendors do already support plain
font files (in TTF and/or OTF format). At the same time all those
commercial fonts have to be distributed as plain font files to be usable
in the operating systems. Nothing prevents the user from putting that
plain file on a public web server... except the copyright law, which
commercial font vendors probably do not *believe* in because they're not
happy with plain font files.


Also see previous thread:
http://lists.w3.org/Archives/Public/www-style/2008Nov/0122.html
Especially the subthread:
http://lists.w3.org/Archives/Public/www-style/2008Nov/0130.html

--
Mikko Rantalainen



signature.asc (204 bytes) Download Attachment
Reply | Threaded
Open this post in threaded view
|

Re: New work on fonts at W3C

François REMY
From: "Mikko Rantalainen" <[hidden email]>
Sent: Tuesday, June 16, 2009 12:35 PM
To: <[hidden email]>
Subject: Re: New work on fonts at W3C

Sylvain Galineau wrote:
> Note that Ascender published a web font proposal last week that
> relates to this charter:
> http://blog.fontembedding.com/post/2009/06/10/New-Web-Fonts-Proposal.aspx

About that proposal...

"However almost all commercial fonts are licensed only for desktop use,
under licenses that do not allow posting to web servers."

There's no technological problem. The only problem is the license for
those commercial fonts! Creating a yet another obfuscated font format
does not change that fact that you still cannot use those fonts because
they are licensed for desktop use only!

>> Copyrighters do so because it's unsecure to allow a font to be used
>> on the web because the font can easily be downloaded and used by
>> anyone who's not authorised.
>>
>> If the format was secured, the problem would (less) occur.

"Commercial font developers are unwilling to allow their fonts, licensed
for use on desktops, to be posted on the web."

Again, this is their choice. Why would creating yet another format
change this a bit? They own the font, they decide how it can be used.


"Most font developers believe that without a technological check-point
(even a simple one), that web developers and server owners will not
understand that they may not simply copy a font from a workstation and
use it on the web."

Please, forward this to tech evangelism department. The key words here
are "most font developers believe". If the font developers believe in
flying spaghetti monster, creating another font format will not help
with that, either. The only real choice is to explain the situation in
terms they can understand. There's no and will not be an effective DRM
system!
>> It's your meaning. Let's other people think otherly.
>> By the way, learn that using the exclamation mark is never seen
>> as a good way to give your meaning as it seems you're closed to
>> other proposal. In fact, you don't listen to them either.

"a technological check-point (even a simple one), that web developers
and server owners will not understand"

Do we really want to endorse any technology which has the key merit of
being too hard to understand to web developers and server owners?

Also note that browsers from multiple vendors do already support plain
font files (in TTF and/or OTF format). At the same time all those
commercial fonts have to be distributed as plain font files to be usable
in the operating systems. Nothing prevents the user from putting that
plain file on a public web server... except the copyright law, which
commercial font vendors probably do not *believe* in because they're not
happy with plain font files.
 
>> Again, you don't look at 'Why do the copyrighters say we can't use
>> the font on the web'... Because "there's no secured way to transmit it"
>> seems to be the key word of their arguments...
>>
>> So, yes, it may be a technological problem, too.

Also see previous thread:
http://lists.w3.org/Archives/Public/www-style/2008Nov/0122.html
Especially the subthread:
http://lists.w3.org/Archives/Public/www-style/2008Nov/0130.html

--
Mikko Rantalainen

1234 ... 14