W3C home > Mailing lists > Public > www-style@w3.org > June 2009

Re: New work on fonts at W3C

From: Tab Atkins Jr. <jackalmage@gmail.com>
Date: Thu, 18 Jun 2009 19:53:30 -0500
Message-ID: <dd0fbad0906181753n690df2dbkfa8257b1ec2bd9d1@mail.gmail.com>
To: "Levantovsky, Vladimir" <Vladimir.Levantovsky@monotypeimaging.com>
Cc: Patrick Garies <pgaries@fastmail.us>, Adam Twardoch <list.adam@twardoch.com>, www-style@w3.org
On Thu, Jun 18, 2009 at 6:05 PM, Levantovsky,
Vladimir<Vladimir.Levantovsky@monotypeimaging.com> wrote:
> On Thursday, June 18, 2009 6:00 PM Tab Atkins Jr. wrote:
>>
>> On Thu, Jun 18, 2009 at 2:58 PM, Levantovsky,
>> Vladimir<Vladimir.Levantovsky@monotypeimaging.com> wrote:
>> > On Wednesday, June 17, 2009 2:56 AM Patrick Garies wrote:
>> >>
>> >> 1. There are free fonts out there and they will likely become more
>> >> prevalent with support for TrueType and OpenType fonts on the Web.
>> >
>> > Yes, it's likely that there will be more free fonts available.
>> > However, the technology like EOT does not preclude the use of free
>> fonts - it will facilitate their use in exactly the same way that it
>> will enable the use of commercial fonts.
>> > The bottom line is that "EOT-like" web font solution:
>> > - empowers web designers to make their own choices of fonts,
>> regardless whether it's free or commercial;
>> > - enables high-quality typography on the web;
>> > - puts the freedom of choice in the hand of the web designer with
>> many different options (both free and for pay) available for them to
>> choose.
>>
>> None of the things that you claim EOT enables have anything to do with
>> technical issues, but rather revolve completely around licensing
>> issues.  In other words, the font format has nothing to do with any of
>> this - if anyone is being limited or restricted it is because the
>> content owners aren't playing ball.
>>
>
> You are right, it is about licensing and I didn’t say that the issues are technical. It has also nothing to do with content owners not playing ball, and we do want to cooperate, very much so - this is why we offer the compression technology we developed - to be used for free and with no strings attached. But as much as I want you and other web designers be our customers - we are in the business of developing fonts for living. So, the final decision depends on many factors such as risks / reward / costs of doing business, which in the end translates into the cost of a font license for you. A solution that is perceived to reduce the risks will reduce the cost of doing business (e.g. font vendors may agree to assume the risks without policing the web) - You will benefit as a results because font licenses will be offered by many different vendors, and at a lower cost.

You are asserting that the issues are technical, however, when you
insist that a technical solution is necessary.  There is no technical
solution to the 'problem' of piracy - it requires a fundamental
rethinking of how you treat content that can be digitized.  This is a
licensing issue, pure and simple, and several of us are saying that it
should *stay* a licensing issue, rather than saddling us with
technical solutions to short-sighted fear of change.

(Amusingly, had talk of obfuscation never entered the conversation -
that is, had the proposal been nothing more than "Hey guys, I've got a
great new way to compress fonts that'd be perfect for web delivery,
and I release patent claims! Wee!" - much of the furor would likely
never have erupted.  The actual merits and costs of the solution would
likely have been discussed much more, without this constant sidelining
into libre vs encumbered formats.)

>> >> 4. Much of the problem with being unable to render certain languages
>> on
>> >> the Web reliably will be mitigated by being able to download fonts—
>> any
>> >> font—with the glyphs necessary to display the characters. This
>> doesn't
>> >> require a particularly high-quality font.
>> >
>> > Web users who rely on downloadable fonts to render certain languages
>> are likely to have slow and unreliable internet connections - they
>> would definitely benefit from downloadable fonts being as small as
>> possible. This is where efficient compression becomes most valuable,
>> and it does not discriminate - it will compress equally well a font
>> that you've got for free, or the one that you chose to pay for.
>>
>> Yup, good compression/subsetting is a valuable feature in any web-font
>> format, because many of the world's non-European descended languages
>> have very large character sets that overlap little to not at all with
>> English characters.
>>
>
> Thank you.

No problem; it's a valid point.  Arial Unicode MS, for example, is a
huge font, clocking in at nearly 23MB.  That's what's required to
deliver good coverage of Unicode, but in most real uses significant
subsetting can be used to drastically reduce the size of the delivered
font, and a well-built compression system may allow browsers to begin
text layout without having downloaded the entire font.

>> > As far as quality is concerned - you are free to choose to use a free
>> font of poor quality or pay a few bucks for a font that is high-
>> quality. It's your website and nobody shall deny you this freedom of
>> choice.
>>
>> There's no need to continue asserting that all free fonts are
>> low-quality.  It's an insult to the work of all libre font designers,
>> and is demonstrably untrue.  While Sturgeon's Law applies as well to
>> fonts as anything else, there is nevertheless quality work among libre
>> fonts.  Especially for headline fonts, which do not require the skill
>> necessary to produce a good body font, the libre font community has
>> produced a number of wonderful specimens.
>>
>
> You are right, there are free fonts available that were produced by skilled professionals, and their work is of high quality.

Cool, then we can stop talking about the relative skill level of libre
vs commercial fonts.  The only relevant detail is the numbers;
commercial fonts currently vastly outweight libre fonts.

(Though, as noted, this can be reasonably attributed (at least
partially) to the fact that there has simply not been an easy way to
deliver fonts over the wire, and so relatively few requests are made
for ones that are free in the appropriate senses.  With the coming
deployment of TTF/OTF support among all the major non-MS browsers,
there may very well be a groundswell of interest.)

>> (It is my experience that basically all of the requests for fancy
>> fonts among designers I have coded pages for are for headlines.  Body
>> text is virtually always acceptable in the standard serif or
>> sans-serif fonts packaged with browsers.)
>>
>
> This is exactly where I see the opportunity for the web to excel - the fonts packaged with the browsers are not likely to support certain (many?) languages. Efficient web font solution would enable any web page be displayed in any of the world's languages - and this requires embedding high-quality body fonts, not just fancy fonts for headlines. Imagine the web where you can access any website in the world and always get to see the content as it was designed, in any language - regardless of the browser or the device you use, and whether fonts you have installed happen to have the glyphs you need to display the text. As a non-native English speaker I can assure you that world-wide web that always "speaks" your native language is a wonderful thing.

In an informal survey of the various computers I own, with a varying
supply of fonts, I can display Thai on a web page with the default
fonts with no problem.  I'm not completely sure if this is drawing
from existing fonts on my system that Firefox is automatically doing
fallback to, or if Thai is simply supported as part of the default
fonts, but I figured an Indic font would be reasonably niche to test
with.

Chinese, obviously, is a bit more of an issue - I know that most
people can't view Chinese on their machines without installing a
Language Pack or otherwise ensuring an appropriate font is installed
on their system.

~TJ
Received on Friday, 19 June 2009 00:54:27 GMT

This archive was generated by hypermail 2.3.1 : Tuesday, 26 March 2013 17:20:18 GMT