¶ Firstly, let me define web font embedding, or so-called web fonts, as I see it. I mean using @font-face to point to regular TrueType or OpenType font files on a web server. In terms of support, the current state of affairs is such that Safari 3.1 supports web fonts, it is scheduled for Firefox 3.1 and it is currently available in a development release of Opera.
It’s true to say that Internet Explorer has supported web fonts since version 4, but only by way of EOT files which are currently proprietary. It’s true that Microsoft is trying to get the W3C to make EOT a standard, but EOT is a form of DRM requiring pre-processing of regular fonts and as such is not acceptable to me, despite Bill Hill’s protestations (the condescending tone and content of which not unreasonably got the Joe Clark treatment).
So. The current state of affairs is that about 3% of browsers currently support web fonts (as I define them), a figure which is due to rise to about 30% probably later this year. This means, to me, that web fonts are not the future, they are the here-and-now, especially if your business is typefaces.
On that basis, it’s high time that font foundries and type designers stopped waving their hands in the air proclaiming the death of their industry, insisting that everyone will be pirating their fonts and installing them for free. Instead they should see this as an opportunity to be grabbed with both hands.
Before I explain how, let’s get a couple of facts straight. Firstly, web font embedding doesn’t install the font on the operating system. The Web Fonts spec has always specified that “downloaded fonts should not be made available to other applications.” So the font exists only within the browser.
Secondly, designers do not necessarily have to upload the font file to their own web server. They can link to a font file on another server. And this is where the real opportunity lies.
When you embed a Google map on your web page, you don’t download a bunch of map images from Google and stick them on your server, you link to Google which then serves up the maps to registered domains. The same approach can be applied to fonts. Font foundries could license their fonts for embedding and serve those fonts only to registered websites, using their own hosted system or via a trusted third party.
This way foundries can provide designers and their readers with a legal way of embedding fonts, removing the need for uploading font files to multiple web servers, and of course make some extra income in the process. Think about it – foundries can sell their fonts twice this way – once to the designer and again to the readers†.
Font embedding won’t increase piracy of fonts – there’s plenty of that already happening via email, BitTorrent and any other Internet vessel you care to think of. Professional designers and reputable clients will continue to license fonts as they always have done. There is no point trying to fight font embedding – that horse has bolted – it’s already happening. So why think of your customers as potential criminals, when instead you could gain control over the situation and make more money from your fonts?
†Update: a clarification and a pricing model
I wouldn’t expect any website visitors to pay for an embedded font. However if designers want their customers to read a certain font then they could be willing to pay for it. So I’m expecting designers (or more likely their clients) to pay for the font embedding on behalf of their readers.
At the moment, fonts are sold on a per-machine basis – a single payment allows perpetual use. But given that the font embedding mechanism I’m proposing is a web service, a monthly payment model could be more appropriate. By way of example, let’s pick a $199 font and divide that figure across two years. That makes about $8 per month, which would seem fairly enticing for something as lovely as say, Archer. And there’s the ‘freemium’ model to think about which might mean you could get the book weight for free, but other weights and italics are paid for. Food for thought, surely?
Comments
1
If I was condescending about anything, which you imply was inevitable, it was not about DRM.
2
See Chris Wilson’s post on this: http://cwilso.com/2008/07/23/fonts-embedding-vs-linking/
Everybody can see that EOT isn’t an effective DRM mechanism; it is not any more effective than ROT13. Because of that, it is really easy to implement. However, for the same reason, it is a total waste of time to implement; if it doesn’t provide any real IP protection, but it exists soley to protect IP, then it is a technological failure.
If everybody cooperates (a big and unrealistic “if”), then EOT does provide a limited form of hotlinking prevention. However, hotlinking prevention is a more general problem that should be solved in a way that is focused on benefiting website owners so that it can actually gain traction.
I don’t understand what you mean by foundries selling fonts to readers. There is no way that enough website visitors are going to pay money for any fonts to make that idea worth pursuing. I’m not sure that it is even economically viable to start selling fonts to website authors. There are very, very few people who are willing to pay even $100.00 for a whole website design, let alone a restricted license to a single typeface.
(P.S. It looks like you need to work on your small caps handling for things like ROT13 and W3C-the numerals are too big.)
3
Joe you’ve misread me. I didn’t mean your response was condescending, I meant Bill Hill’s post was condescending.
4
I’ve added a clarification on that point. I’m expecting designers(or more likely their clients) to pay for the font embedding on behalf of their readers. I wouldn’t expect any website visitors to pay for font embedding, but if designers/clients want their customers to see a certain font then they may well be willing to pay for it.
And yet there are plenty of people willing to buy fonts for installation on their design machines. It is this audience which I am targeting.
5
With support for font embedding in all major browsers, I’d expect companies to end up spending at least as much on web fonts as they do on stock photography, sound effects, etc.
I’d be happy to budget for at least $100–200 (as a one-off cost) for fonts for a typical site, and large corporate clients would pay far more to be able to use the fonts specified in their branding guidelines.
6
My suggestion on Chris’s post was similar to yours: foundries have a way to generate some extra money by offering web-embeddable versions of their fonts. My suggestion differed in that I suggested that font sets may be reduced in some way (no ligatures, no extra weights, etc). The reduced set could easily be marketed as a legitimate way to allow embedding and to get “smaller files fine-tuned for the web”.
7
It baffles me that there is this worry in the industry.
A font is like an image or any other work – if someone puts it on their website without your permission, you sue the arse off them.
8
Why? Worked great for the music industry.
9
I could see this as a selling point on the hosting level; for example say, Media Temple partnering up with a foundry to host a select group of fonts that is rolled up into the monthly fee.
The webhost gets a selling point/premium upgrade to their packages, foundries/type designers get a steady passive pure profit revenue stream and the hosting client gets access to non-system fonts for their sites (I know I would definitely pay for this).
That’s win/win/win across the board.
10
bruce – you’re quite right. What I’m saying is that foundries should provide a legal way for people to embed their fonts, so at least people have a choice of breaking the law or not.
11
I actually think the EOT, assuming it goes through to become a standard, is a much better way to handle things. The Google Maps technology makes a lot of sense living on their servers, it’s a massive application pulling from massive resources, but we’re just talking about typefaces here.
I honestly don’t care too much whether I can buy an EOT version of a typeface or make one from typefaces I already own. The important point is that we find a way to do it that benefits everyone, and also ensures that the work of type designers isn’t easily pirated. Unfortunately, with just a straight up @font-face call for font files the way they exist now, the only way to do that would be to obscure it somehow. EOT is very nice because the file is tied to a URL. This may feel like dirty nasty DRM, but only if the distribution and licensing models make it so.
12
I see where you’re coming from Jason, but it’s already too late for thinking that way. There’s two reasons I say that:
Firstly, the work of type designers is already easily pirated. Any Internet savvy person – including you or me – could get hold of a copy of pretty much any font we wanted without paying for it. I’m certainly not saying that’s right or good, but it is a fact that isn’t going to change any time soon.
Secondly, with Safari 3.1 and the advent of Firefox 3.1, the technology to embed regular font files is already here.
I’m trying to highlight this situation, along with correcting some misconceptions and, importantly, to suggest an alternative which is beneficial to everyone. I really believe a web service is perfectly suited to the job. I’d settle for EOT (if it was supported across the board) rather than nothing, but I don’t see that it’s necessary or useful given the status quo.
13
Right, but embedding normal typefaces without any sort of precautions is putting it on a silver platter for anyone to take freely. That’s like saying people steal from stores anyway, so let’s just leave the doors unlocked at night :)
This alone is a very very bad idea.
The fact that browsers already support font-embedding is great, and if EOT can push through the W3C, I really don’t imagine it would be a long for browsers like Safari to support it (especially since IE, the most problematic case is already on board).
I see where you’re coming from, and I think a hosted solution could be an OK consolation. What I’m getting at is I think it’s better to do it in a smart way, rather than just doing it soon.
14
The font market basically already operates with an ‘honesty box’ approach; it’s pretty easy to find the files for most fonts, but some people pay in order to be legal, reputable and Do The Right Thing.
So I don’t think making it slightly easier to discover the files would make a huge difference, and seems unlikely to outweigh the benefits of creating a whole new market and reaching a new audience (many web designers/developers neglect typography partly due to the lack of font embedding).
Maybe someone’ll work out how to download a gimmicky font and use it in their MySpace page and in a letter to Auntie Mabel, along with some unlicensed photos of kittens & puppies, but so what? Font companies are never going to get money out of such people anyway, but could certainly earn more from companies and web designers.
15
I’m all in favor of some kind of “official” central web font server/service. My thoughts on the matter I think are a less fleshed out version of what is laid out here.
It seems to me that the best way to do web fonts is with some kind of supervision rather than letting anyone use an URL willy nilly.
The google maps example is an interesting model to consider.
16
That’s true to an extent, although it’s not quite the same as directly linking to the font file and advertising its presence. It would still require people to look at the CSS file and work out where the font was situated (not exactly hard, I know, but still requires an active effort to steal the font).
If a service were provided where people don’t need to upload the files to their own server in the first place, then that situation would be reduced. Not eliminated, of course, but reduced. Priced sensibly, I believe many people will choose the legal option.
The thing is, EOT is not the silver bullet you seem to be implying. By being an open standard, it’s inevitable that software will be available to un-encode the font (illegally or otherwise). OK that’s another hurdle to get over but it’s not really much more protection.
17
No absolutely not a silver bullet at all. What I mean is, people who seek to steal fonts can do so now. People who seek to steal embedded fonts will do so too. But the difference I see is a matter of separation. People don’t steal fonts from the source, they steal them from file sharing networks, or something similar. But by making the typeface a simple link in the source of a doc, we are giving it to them plainly, from the source. Meaning, even a pleeb like me could probably write an application to automatically download the typefaces of every site I visit.
After visiting TypeCon a couple weeks back I can honestly say type designers are already terrified of this kind of technology. We not only have to get them to understand that getting their fonts online is a good thing, but also find ways to assure what they’re putting online protects them and their creations.
Yes, having a central spot for people to link to typefaces would work, it could establish trust in an organization to handle the troublesome parts of this. I think that could work, assuming the pricing and licensing get figured out (really the bigger problem than the technology, judging once again from the conversations at TypeCon). I’m still more inclined to host my own files, but whatever gives us the opportunity, and does so in a way that no one gets screwed, I’m all for it.
(Sorry, I’m really not trying to be an ass here! I see your point of view and think it can be really great, I just disagree on giving up on something like EOT just yet.)
18
Of course you’re not. And I’m not trying to be an ass in return either. I do see your point of view too – after all you’ve been there among the type designers, and I agree it’s a hard sell.
I still believe the situation isn’t anywhere near as threatening as type designers believe – to be so would be to imply that vast numbers of previously law-abiding, font-buying designers are going to suddenly turn into thieves. Which, if you think about it, is a rather unpleasant thing to say about your customers.
19
JSM,
While what Rich is proposing here would potentially offer up the fonts “on a silver platter” for casual pirates, it would also be a service of convenience. For the nominal fee spread out over a couple of years (and/or based on per hit/per GB of use), it would be incredibly easy and painless for companies and websites to participate in using the services. This means revenue for the foundries and designers.
Let’s not forget the iTunes Store model. Despite removing DRM for many tracks and Amazon joining the fray with a 100% DRM-free store, it’s more convenient for the average user to buy than it is to steal.
Additionally, the most sought after faces in this web service solution would be fonts that are optimized for Web and not necessarily for print. Print-font pirates wouldn’t care about stealing many of these fonts. I’d never consider using many designed for print faces on the web, I would also never consider designing a print piece using Verdana or Georgia (well, potentially Georgia, but you get my drift).
I see an emerging market for more screen-optimized fonts designed for these type of services. Online we haven’t gone much beyond Verdana, Georgia and the other Office web fonts because there has been no way to feel confident that the fonts we were specifying would be available to a majority of our users. This would make it possible. Suddenly there is a market created where typographers could design to a specific user-base willing to pay for it because of the ease of use (spread-out payments, hosted solutions) and the confidence in end-user support.
20
> you could gain control over the situation and make more money from your fonts
Amen! This is the potential of EOT. I agree with Jason that the barrier it presents, while not insurmountable, is just enough to instill the notion that these specially licensed fonts are not free to be used for any other purpose.
Your pricing model is a good one, but in order for it to really be successful it needs a format that can be tied to a domain. EOT gives foundries the vehicle necessary to market embeddable fonts as a new product, which is what they are.
Even well-meaning, law-abiding font users share and use fonts illegally. A new format keeps things clear: raw fonts have one type of license, EOT fonts have another.
21
This could work.
One question (for now):
> web font embedding doesn’t install the font on the operating system.
What about the browser cache? How hard/easy would it be for somebody (or an automatic utility…) to grab incoming fonts from the cache and put them snugly into the system’s Font folder?
hhp
22
I also support this licence model. But my survey showed that the users don’t like this idea yet:
http://opentype.info/blog/2008/04/19/font-face-survey-results/
23
I’m not sure at the moment how all browsers will be dealing with caching the font file. Safari’s regular cache is fairly obfuscated: you can’t just browse a folder for files, but software is available to pick it apart. It might also be that the font file doesn’t even end up in that cache, but might be retained in memory. Similarly I’m not sure what Firefox or Opera will do with the downloaded font file.
Either way, you’re going to have to make a concerted and deliberate effort to steal the font.
24
How hard/easy would it be for somebody (or an automatic utility…) to grab incoming fonts from the cache and put them snugly into the system’s Font folder?
Not hard at all. I’ve written a short Python script some months ago that does exactly that (well, except installing the fonts into the system automatically). I wouldn’t want to go into details, but it only took a couple of hours to figure it out.
25
Did you see the EULA for Archer? H&FJ doesn’t even allow you to embed their fonts in downloadable PDFs-only PDFs that you send privately to a service bureau can embed them. Further, all other embedding is prohibited explicitly except for Flash. For all other uses, you have to convert the text to outlines or bitmaps.
26
Why don’t font designers charge for every time a font appears in a print ad, magazine, or book? How about in the opening and closing credits of a movie or TV show? What about charging for fonts embedded in Photoshop images shown on the web?
That would be ridiculous… right? Well, guess what: The ONLY difference here is that folks like yourself think charging per use can be pulled off when fonts are run “live” on the web. That’s all. It’s not a question of “fairness”, since the other uses I mentioned are NOT charged. It’s a matter of what can be gotten away with.
What will actually happen is – no one will pay for fonts used this way. They’ll either not use fonts that cost extra (i.e., stay with the status quo), or if they really wish to be different, they’ll opt for fonts that waive any usage fees.
If fonts designers want to make money, the way to do it is to allow their fonts to be used online at no extra charge beyond the cost of the font, whatever that is. All that should be required (and this probably CAN be an automated process) is some sort of user ID/serial number. Just like Photoshop or any other software program, fonts won’t “run” unless the website owner owns the font.
THAT is the business model that wins out here, not a charge per view or rental model. And the marketplace will sort it out, as the font designers who grok this and offer this pricing plan will far outsell their competition.
27
I’m not sure quite sure that will fly, but I won’t be complaining if you’re right.
I see Mister Snitch’s point about web just being another use, just as if using a font in a TV show, magazine or jpeg, but in fact many font licenses are quite specific about the medium in which they can and can’t be used.
28
This is great news. I wasn’t expecting something like this for a while, as I thought there were more important things to attend to. IN M$’s case, getting the browser to have some form of consistency across versions.
29
The cache accessibility issue has to be resolved, otherwise any judicious font house would rather be safe than sorry.
> Either way, you’re going to have to make a
> concerted and deliberate effort to steal the font.
The point is, does it take 5 seconds, or 5 hours? Except for people whose principle it is never to pay for fonts (or software in general) this is critical, because it makes the illegal acquisition of a font cost-effective or not. Just like extracting from a PDF: yes you can do it, but you might need to spend a few hours tweaking the results, at which point it’s smarter to simply pay for the thing.
It is not just an ethical issue; to a much larger extent it’s a practical one.
I’d love to believe that this can work, but like if Jens is right, this thing is dead in the water.
hhp
30
Why don’t font designers charge for every time a font appears in a print ad, magazine, or book? How about in the opening and closing credits of a movie or tv show?
Seems like you have a rather negative opinion about font vendors, haven’t you?
Well, I am selling fonts and when I think about webfont service with a rental model I do this because I think it can be more flexible for the users and maybe even cheaper than every previous license system.
I agree, if you buy a print font, that might also include a web license in the future. But what if you want to buy a font for your website only? Do you really want the old system where you have to pay the full price before even getting the chance to try the font. Wouln’t it be better to license a font for $1/month and try it out for a while? Maybe even try different fonts for that same price?
I’d love to believe that this can work, but like if Jens is right, this thing is dead in the water.
Webfonts will increase font piracy. There is no doubt about that. No matter if someone reverse-engineers a EOT font or picks the linked webfont out of the cache. It will always be possible to steal webfonts. But that’s not the point. The question is, if the vendors will loose or win in this game. When hundrets of websites would use a font without a proper license it would be bad. But if at the same time the same font had thousands of customers who pay for it, then the foundry should still offer it. That’s how it works for the music industry today. Even though you can get every song for free, there are still enough people who buy the songs legally.
That doesn’t mean we don’t need to protect the fonts. It still needs some degree of protection. But that’s a secondary issue. We should first think about a license model that works for the user the foundry and think about the security issues afterwards.
31
Ralf, I certainly agree that allowing people to steal our fonts is OK as long as we’re selling more than before as a result of the new sales channel. But I don’t see how security can be secondary. I think in this scheme we already have a new licensing model that can work, but ONLY if it’s non-trivial to initiate the piracy of a given font. If all a person would have to do is open a cache folder and copy the font out (in ANY browser really) forget it.
BTW, that survey you did was awesome. Thank you.
hhp
32
If all a person would have to do is open a cache folder and copy the font out (in any browser really) forget it.
Not sure how this will play out in Firefox or Opera, but for Safari you can already build the font in a way, where the file ending up in the cache can not be installed in the system.
33
What if I wanted to keep the website going for more than two years – would I have to keep paying the monthly fee, and end up paying much more than an outright license for the font? And if not, if I am in effect only paying off the value of the font, isn’t that just buying a license – which there seems to be opposition to?
34
I think the rental model offers many advantages for the user. But maybe it would be a good idea to give the users a choice.
Almost the same way as buying a car:
1. I know what I want, so I buy the whole thing and use it as long as I please.
2. I lease it for some time and “give it back” when I’m done
3 I lease it for some time and when I want to keep it, I pay the rest of the price (if there is any) and can use it then for as long as I wish.
35
It is good to read this as a positive endorsment for using special fonts now. One of the desginers I work with is a big typography nut and wants to use some special fonts on the site which I have discouraged. I suppose I may have to change that answer in the coming months.
36
Here’s a wrinkle in the conversation:
Cory Mawhorter has apparently built a JavaScript called “Facelift” http://facelift.mawhorter.net/
The technique is billed as an alternative to “sIFR”, and builds images on the fly for replacement of headings/text.
It apparently requires the font file to be uploaded to a folder on the server.
The blog & forums don’t mention the font licenses yet.
Add your comment
Comments are now closed on this post. If you have more to say please contact me directly.