[Below is the manuscript of my presentation at Atypi 2008 on Sep 20th in St Petersburg, Russia. The presentation was part of a panel where Thomas Phinney, Simon Daniels, Bert Bos, Roger Black (chair), and Håkon Wium Lie (me) participated.]

Web fonts: the view from the free world

I’m a web guy. I want everything to be available on the web. I think everything will be available on the web. I think the web will hold the master copy for all of mankind’s data and documents. If you want off-line presentations of the content, you will still produce it from the web. Along with Bert Bos, who is also on this panel, I wrote a book that showed the feasibility of this approach; the book was written in HTML and CSS and converted to PDF for printing purposes. In the future, I believe all books will be printed off the web. If so, the web must embrace a new data type: fonts.

CSS, in itself, is also an example of a new data type on the web. Browsers before CSS had style sheets, but they were hard-coded. CSS style sheets, on the other hand, are web resources that can be trasferred, referred to and reused.

CSS is now, after a long struggle, interoperably supported in browsers. We now have a set of finely tuned font properties in a billion browsers that can be used to generate pixel-perfect web presentations. These fine typographic machines, however, have few fonts to process. There are only about 10 fonts that can be reliably used on web pages. Microsoft made the fonts available to the web community in the 1990s. They have served us well, but have now become a limitation; there's only ten of them, and the filed cannot legally be changed. When designers want to do something fancy with CSS and fonts, they resort to background images instead of using fonts. Background images! That’s actually how most designs in the CSS Zen Garden are made.

Therefore, a few years ago, I started a campaign to make browsers support webfonts. In order for webfonts to succeed, browsers would have to agree on a common font format and a method for linking to fonts on the web. Now, a few years later, we have four implementations that support Truetype (or, OpenType, as revised edition is called) fonts through CSS style sheets. There is one shipping browser (Safari), two experimental browser implementations (Opera and Firefox) and one web-to-PDF-by-way-of-CSS converter (Prince).

Using these implementations one can create some truly marvelous web documents:


There is not a single image on those pages!

These test pages use fonts that are freely available on the web — there's a bunch of fonts to choose from. SIL has published Gentium in a webfont-friendly license. James Arboghast has made a subsetted version of PykesPeak, called PykesPeakZero available under another liberal license for use and reuse on the web.

Will there be a commercial market for webfonts? I think so. Many web sites will want to have custom-made fonts on their sites.

Will we want to use webfonts for all text on the web? Probably not, for body text, you may want to continue using a locally installed fonts. For display type, the opportunities are endless.

There’s an ongoing discussion about which format to use for webfonts. Microsoft has submitted the EOT format to W3C. EOT has been supported for years in IE without anyone noticing. I’m glad to see that the format now is published. However, that doesn’t automatically make EOT suitable for use on the web.

EOT is an obfuscated file format built on top of OpenType. There has been some discussion whether the extra layer is DRM or not. Certainly, EOT is digital, it says something about the «rights» to the fonts, and it mandates management of those rights. Still, DRM or not is the wrong question to ask. The right question to ask is whether EOT falls under DMCA. Can you be sued for «breaking» the «protection» in the EOT file? From the outside, it seems that one of the few remaining reasons for pushing EOT (as opposed to Truetype) on the web is make sure that DMCA can be used for legal action.

If DMCA applies, it could turn browsers into circumvention devices; browsers would have to decode and render the «protected content» and could, possibly, be used to circumvent the «protection». If someone would like to make a legally informed statement about browsers and EOT, I’ll be listening.

If a case is brought against someone, being innocent doesn’t keep you out of trouble. I want to share with you a story from my own neighborhood. I have a good friend called Jon. The world knows him better as «DVD-Jon». As a young kid, the Norwegian police raided his family’s house one early morning. Jon had published code that allowed people to play DVDs on Linux. Thereafter, Jon spent years in the courts. While these cases made their way through the legal system, some of us took to the streets of Oslo to give Jon support. Do you remember what the DVD encryption system was called? Content Scramble System — CSS! So, in 2002 I found myself chanting anti-CSS slogans in the streets of Oslo. Here are some pictures from the 2003 May day parade in Oslo where it is customary for people of various political flavors to make their views known:

Jon prevailed in two rounds in the courts. He did remarkably well and I think he learnt some things that may come in handy later in his life. Still, I don’t want for others to face the same ordeals for trying to decode webfonts.

Then there was this other person, Dmitry Sklyarov... no, I think I’ll stop there — we’re getting a bit too close to home.

W3C is currently faced with a difficult question: should they bless the EOT file format when Microsoft asks them to, or should they listen to the other browser vendors who have reached consensus on TrueType? If W3C does not bless EOT, the format is dead and Microsoft is likely to join the other browsers in supporting TrueType on the web. This, in my opinion, would be the best solution. However, if W3C decides to bless EOT, we could face a long-standing format war on web fonts. Authors will have to provide fonts in both TrueType and EOT and they may chose the easy way out, which is to do neither. I certainly hope that W3C will not do something to damage the consensus we have have worked long and hard to reach.

So, dear typographers, I invite you to come along and start using webfonts. And to start selling webfonts. Make some fonts that are worth the download, fonts that are interesting, fonts that are inspiring, fonts that make us laugh, and fonts that are beautiful.

However, if you choose to stay behind the iron curtain — the lead curtain perhaps — along with Microsoft and Adobe, you are welcome to do so. The free world will not bother you, and we will not use your fonts on the web. We will find other fonts to use.

It’s important to understand who is making the final decision about webfonts. It’s not me. It’s not the people in the panel. It’s not W3C. And it’s not you in the audience. The decision is made by the people out there who create web pages and style sheets. If they like what they see, webfonts will soon be changing the face of the web.