Whoops. Here's the blog post where those comments are happening:
http://nicewebtype.com/notes/2009/07/19/type-sellers-web-fonts-and-typekit/

I don’t see people in those comments arguing that fonts should not be licensed; they’re just pointing out the obvious; DRM does not prevent piracy, DRM has historically been a failure, and there is no reason for a font DRM system to be put in place on the web that trumps all of the reasons not to.

You're right James, thanks. I've rephrased the forum topic. Two questions now: licensing vs. DRM, licensing vs. freebies.

Fonts can be re-used in unlimited ways, making them more like a tool, whereas other web content is typically not seen as valuable for re-use. All the applications for fonts, from books and magazines to identities and websites, also have potential for their creators to make money. You can try, but you can't legally make money selling photos, writings, movies or music that you aren't author of, or legally entitled to sell. Fonts have similar value as tools, but the new situation presents the possibility of those tools, which still have value, to be distributed freely, against the will of the original designers.

DRM and everything concerning it makes me think of the method a friend of mine devised to make smoking cigarettes a thing of the past: people are encouraged to stop buying cigarettes but instead boost them off friends, acquintances and passersby. Who after doling out their cigs for a while will stop buying them, because it’s easier and cheaper to boost them off other people. The end of the story is that there will be just one person who has to buy and pay for all of the smokes. And when he/she decides to stop doing that, the problem will be solved.

Sounds like a great plan, right? Yeah, it’s just as if type designers will let everyone use the products of their effort and time for what? A pat on the back? A cigarette?

What piracy means is that the honest people who pay for the product subsidise the pirates. And as soon as these honest johns turn in to blackjacks and stop doing that, there will be no more for [insert product of choice] for anyone. And alternative marketing methods — as suggested on ILT — will not make up for that, or counter the trend.

BTW My friend made national television with his plan (in the Netherlands, okay…).

. . .
Bert Vanderveen BNO

crossgrove: Um, graphic designers and photographers would definitely have a problem with what you wrote, as they find their work can get copied and used on other web sites and even moved to print, that others make money off. It's all just intellectual property, that is "owned" by an individual/company and licensed to others with various limitations in the license.

Could a font be used on every web page on the planet? Sure. Is it likely to happen? No.

Will some web sites, either accidentally or intentionally, use a font for which they aren't licensed on their web site? Definitely. Will DRM make a significant difference? From looking at how other IP is treated on the internet, it is unlikely.

People who professionally make web sites (either inhouse or freelance), or care about their brand/image, will make sure their font usage is properly licensed. The ones that don't, or individuals just making one-off pages, are less likely to use web-based fonts because it costs more to host/deliver their web site, and even if that doesn't affect them, then they still wouldn't pay the licensing fee, even if asked (at best, they would simply stop using the font).

The foundries with complicated licensing agreements will find that they simply are impeding their fonts from being used legitimately, rather than "protecting" their font. Adding DRM which will take years to adopt just keeps them in a bigger hole than other foundries without these requirements.

bert: "What piracy means is that the honest people who pay for the product subsidise the pirates. And as soon as these honest johns turn in to blackjacks and stop doing that, there will be no more for [insert product of choice] for anyone. And alternative marketing methods — as suggested on ILT — will not make up for that, or counter the trend."

Um, the honest people are NOT subsidizing the cost for pirates. There is no actual physical cost to the font producer for someone that hosts/uses a font without the proper license.

If there are not enough people who want to license your product (font, graphic design, photograph, text, whatever), then there is something wrong with your business model. You are charging to much, it's too difficult to use, you need to advertise or all kinds of other things that affect that business model.

History has shown that DRM harms an industry, rather than help it (see music industry for example number 1).

Cost is cost. Whether it is the cost of the Chinese girl who assembles your iPhone (at a mere 30 USD a month) or the typedesigner who designs the fonts on your business card. Paying the Chinese girl is a nonbrainer — even if you steal an iPhone, she gets paid. The loss will be distributed to all of the other buyers.

Paying the typedesigner should differ from this model, because the product is bits and bytes and there is no ‘physical cost’? Maybe… But in that case you should accept that new business models will evolve with the very same result as the original one: the end user pays. Upfront or down the line, just as it has always been.

. . .
Bert Vanderveen BNO

bert: The idea that every use of unlicensed IP counts as though it were physical theft, and that all the people who do license it need to pay for those unlicensed uses, well, is ridiculous. I know that's how groups like the BSA like to calculate so-called software industry losses, but that doesn't make it a reasonable argument.

You seem to be equating font designers getting paid with DRM. Other industries have shown that DRM winds up hindering, not helping those industries.

The more obstacles you put between yourself and your customer (wacky licensing agreements, drm, etc...), the more likely your customer will be to find another solution (that may or may not include pirating your font).

There are lots and lots of web designers and companies that make web sites and make sure all the elements on their pages are properly licensed (graphics, photographs, text) without needing browsers to have DRM embedded in them. If using a specific font for a web site is perceived to have enough value (ie, the same or more value than it's licensing fee), a designer will license and use it.

If there are enough designers who value the fonts you provide enough to use them (as in, the value they add to the site meets or exceeds the amount you charge), you have a viable business. If there aren't, you don't. Adding a layer of DRM just increases the 'charge' side of the equation, without any benefit (as it costs everybody, font designer needs to configure and add it to the fonts, along with any infrastructure it needs, web browser needs to implement it, paying customer needs to figure out how to make it work with their web site).

Let me get back to this tomorrow — it’s kind a late over here and after the single malt something else needs attention. . .

Bert Vanderveen BNO

Watch List

I’m not with DRM or something, but let me see if I get it right:

Let’s say that you, as a type designer, need 100 dollars a month to live. If 10 designers are paying for your fonts, you can charge 10 dollars each.

But if only two of those designers are paying, the easiest solution is to charge 50 dollars each. So the honest designer is paying 50 dollars instead of 10. In this sense, I’d say that the honest designer is paying for the pirate (subsidizing?). Another option: you can start selling cheaper, hoping to make more sales (but no guarantee).

Now, if at some point those two remaining designers stop paying (“Why should I, if I get the same stuff by free and no one else is taking the right way?”, “Why should I pay 50 bucks for this, if it should be costing only 10?”), you need to find another job and the world won’t see more fonts by you (or anyone else who takes this job seriously, given the effort it implies). But your type designs are helping those 10 designers to make money, right? How is that? I don’t know if piracy is the only responsible for this, but definitively doesn’t help.

Now, the big problem is how you can convince the non-paying guy to change his mind, but it’s another story (and the original question of this thread, I am afraid).

Am I right, or am I missing something?

Cristobal: "Am I right, or am I missing something?"

You're missing that it is violates copyright law.

Sure, somebody may create a blog on their own host that may use your font, and blow off your complaint to them about it, but any non-trivial company (which is what virtual all web-designers are doing work for) will care about their public-facing web site blatantly violating copyright. They won't like that the web designer they paid to make their web site opened them up to a lawsuit, at least potentially costing them thousands of dollars in legal fees, let alone settlement fees.

Blogging sites like wordpress and sites like twitter/myspace/facebook will probably want to offer new fonts as a feature, and they'll make sure to get a license for those fonts (and just not permit the use of fonts from the foundries that attach unreasonable licenses/fees for the fonts). And they probably won't let users upload random fonts.

Ok, if I get the point, your answer to “Why must type be protected?” is not be lawsuited. Good reason, indeed.

In your experience, does it work? At least in Latin America it is not the case. As a typographer, I’ve seen that many companies don’t have licenses for their fonts, but I’ve never heard about a company (trivial or not) being lawsuited by this (I’ve heard about lawsuits for using other kind of illegal software, but certainly not fonts).

But that can be just because I live in the third world. Maybe someone there outside can name foundries/type designers following this lawsuit-strategy and what the results are/were (I recall vaguely a thread talking about this, is it possible?).

Cristobal: Your countrymen don't have a high regard for IP. Being in a business that produces IP and trying to sell it locally may not be the smartest thing to do. Slapping on a layer of DRM on fonts won't make a difference to them. If they want to use it on their web site, they'll do it, and it'll probably be easier for them than for your paying customers because they won't have to worry about making sure the DRM works right when somebody visits their web site.

However, other countries, with perhaps a lower piracy rate, where lots of companies spend a lot of time and money on branding and image, they do take IP seriously. And they have the money and are willing to spend it to get IP they want.

And IMHO, piracy for web fonts could easily be less than it is for computer fonts, just because they will be public. These fonts are not aimed at being used on internal wikis. They are for creating a more compelling public web site for customers to see. Foundries can readily find out if a web site is using one of their fonts. With computer fonts, it's much harder (or impossible) for a foundry to tell that a given company is using their fonts for all their correspondence, because where they are being used is internal-only, or just to clients (or even in printed matter).

Lots of companies (in the US at least) will do the right thing (either stop using the font or pay to license it) if notified they shouldn't be using it.

Simply put, for a business that sells IP of any kind, a strategy of aiming your sales strategy at people who have money and are willing to part with it in exchange for the IP works much better than trying to shoot arrows into everybody that will never in a million years pay for your IP.

My best answer to the people who ask why typefaces should be protected is to ask that person if he would mind if I cashed his paycheck this week. It can help to understand that fonts don't occur in nature. People create them, and while some might create them freely, and offer them to the world, most of the best ones are created by people who put in huge amounts of labor and rather hope to get some recompense.

I'm tracking this thread.

Dave, I don't think I see anyone in this thread arguing for DRM. Protection doesn't have to come in the form of DRM; it can come in the way of a very well-designed distribution model, and that's what's being discussed.

Okay, DRM may not work in the longer term because it is obstructive to the user, but at least it heightens awareness of the fact that (music, movies and) fonts are IP and cant be used, enjoyed and distributed freely.

Certain countries should be considered to be developing countries. Eg China, Rusland (and apparantly large tracts of South America), but as soon as these develop enough to have their own creative cadres in place, legislation will adapt and IP will be protected there as well.

(Image being a Chinese typedesigner who has done a popular font and sees it in use everywhere without being properly compensated for drawing these 1500+ glyphs. He’s gonna blow and lobby like hell in Bejing, right? It takes only one of those with an uncle in the government…)

. . .
Bert Vanderveen BNO

As Frank Martinez pointed out during the prep meeting for the TypeCon web fonts panel, DRM is a business model not a technology. Confusion over this is one of the reasons why we decided to avoid the term during the panel discussion.

No one that I know of is advocating a web font format solution that would enable a DRM business model. It simply isn't on the table because unlike, say, iTunes music, a web font format needs to be supportable directly within browsers, not in special web font 'players' or via plug-ins, and most of the browser makers will not implement a format that contains technical measures to protect software that require enforcement on the part of the browser. What some of us are advocating is a format that a) is appropriate to the web medium (which means a format including compression) and b) contains bulkhead protections against casual and unintentional unlicensed use. The difference between the bulkheads and the kind of technical measure that would enable DRM -- the kind of measure that automatically invokes DCMA law in the US -- is that the bulkheads simply sit between the web font and casual breaking of license agreements, requiring no special action on the part of the browser.

DrDoc, this entire thread is about DRM, as a) that's whats in the title of this discussion, and b) that's why the variety of new font formats are being proposed for. The existing font formats would work fine with font-face, the big foundries don't want to use them for the web because they don't have DRM and can be readily downloaded and used both on other web sites and on your local computer.

The problem with this is, either the format conversion from say truetype or opentype to a new 'webfontformatwithdrm', is that either the conversion is lossy (as in, the web font is incomplete compared with the original, reducing the usability of the font), or there WILL be conversion tools that reverse the process (and strip out any drm).

Thanks John, for that clarification.

I think that the DRM vs free dichotomy that many keep asserting in these threads is really a false dichotomy. There are more options than either trying to make it impossible to steal a font, or to make it free. Two things can I think and hope create a sustainable market for web fonts: 1. Making it not easy and obvious to download the font for non-browser use without piracy software; and 2. Enforcing large, public violations of EULAs.

Because fonts are tools for publishing either on the web or print, and because publishing is by definition public--and the web especially public--the situation is significantly different from the story with music. Font vendors are not going to even attempt to go after the individual consumer, which was the issue with downloading music. But they will go after publishers.

And I think large foundries such as Monotype, Adobe, will be willing to sue owners of web sites who do considerable business for violating EULAs for web fonts. And once the principle is established, my guess is that web businesses will be extremely reluctant to do stuff in public that makes them subject to a hefty law suit, when there is a relatively low cost to getting the font.

So I think that while significant control of piracy among say scrap-bookers may be impossible, it does seem possible to have enough enforcement leverage on web publishers to create a remunerative market for web fonts.

@John Hudson

Browsers these days support general gzip compression for all content loaded via the HTTP protocol, including IE. It's unfortunate that the Typecon web fonts panel gave the distinct impression to everyone attending that a new file format is somehow required to get compression of fonts, it is not. Rather the question is really would font-specific compression techniques provide additional compression that would justify the additional code required to support it. In many ways, I think supporting better general compression techniques would be a far bigger win for the web. For example, general lzma compression appears to do a better job compressing your Constantia font.

What precisely is a "bulkhead protection" against casual font use? Are you talking about a wrapper format that contains obfuscated font data or a format that uses some form of automated licensing information to regulate font usage?

John Daggett
Mozilla Japan

John Daggett, my impression was that the panelists agreed that built-in restrictions on what web site the font could be used with would not be accepted by any browser outside IE, because of potential liability issues. The proposals discussed by the panel were "wrappers", and, as you say, font compression.

@John Daggett

Thank you for your clarification regarding compression. I apologize if the panel gave the wrong impression that font-specific compression within a unique format was the *only* way to improve file size and performance.

I thought that Shu's response to the compression issue addressed that. But perhaps it was too fleeting and non-specific to fully communicate the point.

I'd like to think it isn't an either/or proposition and that any additional compression both inside and outside a font format would be worth some extra code. I happen to live in a rural area, underserved when it comes to broadband access. I finally just this year upgraded from dial-up to satellite -- that's all that's available. To me, personally, every little bit makes a big difference.

-- Kent.

To the original post,

>Why must type be protected?

Type, must not be protected, ever. Fonts must be protectable though. Seeing this from the user perspective, it is not entirely totally the opposite of other media, but close.

>Why do you believe fonts are “special” for the web [...] vs any other licensed media such as graphics, music or video?

Type must be protected. (just kidding). Fonts, not so much special, are different because the operating system's treat them differently.

>[How are type designers different] than the creator of images on a webpage, or sound, or text?

We're all human, beyond that what's the question?

Kent, I think the panel gave entirely the wrong idea of compression. David or someone jumped raw and naked at Bill or someone when they implied that compression is a bad thing and nobody stopped him, nobody. Then David said, compression is not a bad thing... but well, if you don't have to compress a font, that's a good thing isn't it? Compression in this instance particularly sucks because digital outline fonts are so friggin compressed, you couldn't fit a credit card in there. What's more, there is danger in font compression that makes, e.g. your connection speed issue trivial. So, maybe compression is a separate and slowly cooling issue ;). Subsetting, now that should be a hot issue!

The only other thing I might have bit on panel-wise, and maybe other people thought of it too, was when the user asked from the audience what would happen if, (though it's more likely when), their @font-face call doesn't work: the panel agreeing unanimously for a change, that the backup font would be there, probably like Verdana... But what if text is replacing a graphic of text, like everybody says they are trying to with @f-f? Can web developers write code to, "Either display this font, or if you can't, tell me, and I'll tell you to display this graphic instead?" It might be trivial, but then it is the web we're talking about so my guess is that there are going to be at least some size differences to cope with, getting from a bunch of scaleable letters back to a block of pixels of text... I mean no one wants their logo e.g. to default to Verdana or anything else, do they?

Cheers!

@kentlew

I’d like to think it isn’t an either/or proposition and that any additional compression both inside and outside a font format would be worth some extra code.

This is not necessarily true, running gzip over a file twice generally won't result in a more compressed file, just as enclosing a JPEG in a zipfile won't reduce the size. The EOT format used MTX compression which includes two different types of compression, Truetype-oriented compression of things like hints and metrics and general entropy compression similar to gzip for everything else. [1]. With heavily-hinted TrueType fonts, MTX does much better than gzip but with more lightly-hinted fonts (e.g. Microsoft Cleartype fonts) the results are more marginal [2].

Also unmentioned during the panel discussion was the fact that the embedded font API's on Windows used to support EOT do not support Postscript CFF fonts, so there's no backwards compatible way of supporting these fonts in Internet Explorer, other than to convert them manually to fonts with TrueType glyphs.

[1] MTX description
[2] Comparison of MTX compression to gzip and lzma

@dberlow

Compression in this instance particularly sucks because digital outline fonts are so friggin compressed, you couldn’t fit a credit card in there.

Sorry but this is completely untrue (see chart). For example, general gzip compression will reduce Constantia from 314K to 154K, less than half its original size.

Subsetting, now that should be a hot issue!

Agreed.

Chart showing compression of common Western fonts

> This is not necessarily true, running gzip over a file twice generally won’t result in a more compressed file, just as enclosing a JPEG in a zipfile won’t reduce the size.

Fair enough. And yet we don't find TIFF as a graphic standard for the web, relying on server-side gzip compression to handle size issues. Instead we prefer already compressed file formats like JPEG.

> Kent, I think the panel gave entirely the wrong idea of [ . . . ]

Oh well. Ain't live performance fun!

And yet we don’t find TIFF as a graphic standard for the web, relying on server-side gzip compression to handle size issues. Instead we prefer already compressed file formats like JPEG.

Poor analogy, TIFF as commonly used is a lossless form of compression, JPEG is a lossy form which results in dramatic compression ratios. All the font compression techniques discussed are lossless (gzip, lzma, MTX) so the question is how much better does a given compression method improve compression ratios versus already available methods (i.e. HTTP gzip compression).

As David Berlow suggests, a "lossy" form of compression like subsetting is where substantial gains are to be made in reducing the size of fonts served. This doesn't require a new font format, it requires new tools that do this accurately without quality loss. Typekit is effectively providing a tool like this, a tool used to create a custom font for each user of their service.

What’s all the discussion about fontfilesizes about? Everyone in the richer parts of our world has bandwidth to spare and no one cares about the rest.
Just watch Activities in Safari and notice that what swoozes by every time you hit a new webpage is the same amount of data that a layout in QXP was in 1989.
Forget compression. Talk protection.

. . .
Bert Vanderveen BNO

Also unmentioned during the panel discussion was the fact that the embedded font API’s on Windows used to support EOT do not support Postscript CFF fonts, so there’s no backwards compatible way of supporting these fonts in Internet Explorer, other than to convert them manually to fonts with TrueType glyphs.

Unless MS decide to interpret that as a bug, rather than as an unsupported feature, in which case a hotfix is possible. Given that it is in their interest to foster a format whose principle benefit is being backwards compatible, I would expect MS to do just that.

What’s all the discussion about fontfilesizes about? Everyone in the richer parts of our world has bandwidth to spare and no one cares about the rest.

One billion Chinese, one hundred million Japanese and fifty million Koreans beg to differ.

Which begs the question if the current approach of 1:1 match between character and glyph is so ideal for these scripts. And also begs the question if fonts, even those covering only Latin, Greek or Cyrillic scripts, can save a number of glyphs by making heavier use of mark/mkmk feature (which however requires that layout engines support the OpenType specification more than piecemeal).
Off-topic. At least as much as is discussing file size and compression in the context of protection.  :)

Remember, we're just trying to get a simple web font format which offers some protection. EOT Lite offers that. Any number of other proposed formats would be adequate as well. The persistent use of the simplistic and inflammatory term "DRM" is really not helpful to the discussion. EOT Lite offers enough protection for many foundries and has nothing even vaguely resembling DRM in it.

One other thing to consider: The DVD format offers relatively mild copy protection, and it did rather well. You could say that there was a small but vocal opposition to it being there, but by and large the average customer didn't even notice. It worked, and the format was one of the most successful new technologies in history. You might call that apples and oranges, but it's an example of an industry having enough protection to be comfortable putting its digital content out in the world without causing any particular inconvenience to customers.

Adobe does have reason to expect that recent and future versions of IE would support CFF fonts in EOT. Possibly older versions as well, with a little help from Microsoft. With the compression inherent in CFF, the compression issue becomes less urgent.

And yes, subsetting can be pretty handy.

South Korea has the BEST data infrastucture in the world… Japan likewise. China could have, but they throttle and block, so it’s upon themselves.

. . .
Bert Vanderveen BNO

What’s all the discussion about fontfilesizes about?

You'll know when you hit a page that e.g. links to FF Trixie HD Pro (file size 16 MB). And for services like the one by Typotheque, who disallow caching, the font file would have to be retransferred for every single page you load.

South Korea has the BEST data infrastucture in the world… Japan likewise. China could have, but they throttle and block, so it’s upon themselves.

Well, good for them, but in some of the more rural areas of Germany for example, you still don't get anything over 64 kbit/s.

dberlow>Compression in this instance particularly sucks because digital outline fonts are so friggin compressed
jdagget>Sorry but this is completely untrue

Your chart's too big. Verdana is 205k complete and (external) Compression can get it down to 100?
Subsetting, e.g. for cap setting, gets it down to 36k or less with no compression, and still setting all sizes that may be required.
Can we see a light-table comparison of the original drawings of Helvetica and EOT Arial output at 936 point 6000 dpi please. ;)

I've been through this at and after a standards meeting with several of these folks here. I think I can guess what most technologists want to do, and when they are finally finished, compression will no longer matter for some other unrelated reason, and if subsetting has not supplanted compression anyway, it'll be a bad sine.

Compression is bad, if it gives a false sense of security to font owners, if one wants to avoid end user decompression, if it's the only solution available to font file size reduction, or if unknowingly to the web developer and/or apparently to the user, it degrades functionality or quality of type.

Chers!

> Well, good for them, but in some of the more rural areas of Germany for example, you still don’t get anything over 64 kbit/s.

And rural Massachusetts, USA. ;-)

Well, good for them, but in some of the more rural areas of Germany for example, you still don’t get anything over 64 kbit/s.

And rural Massachusetts, USA. ;-)

and rural Oklahoma...
~~~~~~~~~~~~~

...from the Fontry

And rural Massachusetts, USA. ;-)

and rural Oklahoma...

So at least Germany is not the only third-world country when it comes to Internet access :)

@Dave Person: he existing font formats would work fine with font-face, the big foundries don’t want to use them for the web because they don’t have DRM....

Most foundries are not at all big, and the desire for some degree of protection (which may or may not be "DRM" depending on one's definition thereof) is voiced as frequently by individual designers and tiny foundries as by large ones..

Many people who know nothing about this business try to blame anything they don't like on big business. In this case at least, the size of the business is unrelated. The last western type foundry to do serious DRM was Letterhead Fonts, who are hardly a "big foundry."

Cheers,

T

The desire for protection is also voiced by clients who commission custom typefaces and wish to maintain the value of their brand.

i can certainly understand "the desire for protection" and
recognize that it can come from sectors across the map...

i can also understand a desire for perpetual motion.

and me personally, i'd like to fly through the air just like
i swim through a pool, with my nice solid breaststroke...

don't listen to those people who say "impossible!'

-bowerbird

p.s. i live in santa monica, which is hardly a rural setting,
and sometimes i can't get an at&t cell signal in my house.

>sometimes i can’t get an at&t cell signal in my house

Your house is a cellular device? Are you small, or is it a big device?

david said:
> Are you small, or is it a big device?

i'm phat, and the apartment is a 2-bedroom-1-bathroom device.

-bowerbird

while we're in the midst of straightening out this whole font thing
i have a question...

as for the individual posts on this here typophile forum thingee,
is the font set by the writer, the reader, or the .html programmer?

i wanna be able to set the font to comic sans, maybe papyrus...

but only for my friends!

for enemies, 4-point grey dingbat please, with negative leading...

-bowerbird

Re: Compression
As John Daggett points out, there are substantial reductions in file size to be had over CFF alone.
And the numbers for LZMA really caught my attention.
At the server level, gzip reduces the size of the file as it's transmitted with the browser un-gzipping it on the other end.
However, I have a problem with saying "use gzip" and letting it go at that, because it's exclusionary - it assumes a level of technical expertise and control over the web server. Compression built into the file can be created by anyone even if they don't know a .htaccess file from a hole in the wall. Compression at the file level, is just simpler. And I believe, as a rule, that the closer web development stays to a do-it-yourself proposition, the more inclusive and democratic it will be. I don't think JD or anybody will argue differently.

@tim brown
As to the original question: as with any grant of copyright, granting the monopoly only makes sense if it helps to promote the "useful art" (as the US Constitution phrases it). In other words, if it helps keep the work product we need coming at a reasonable price. In the case of fonts, which are both labor and knowledge intensive, I believe it does. Others disagree.

@bb
You entertain me, as_always.