FontSlice: Distribution Evolved

[David Jones]

This topic was imported from the Typophile platform

Starting a new company is a long and painful process, but we're finally ready to show off the fruits of our labor. We hope that FontSlice will fundamentally change the way fonts are bought and sold on the Internet. Please email me at [email protected] if you're interested in learning more about a distribution agreement. Now, without further ado, meet the iTunes of typography: https://typography.guru/forums/topic/101389-forwarding.

[Ryan Maelhorn]

so umm, maybe you should tell us why and how this will revolutionize font sales on the web? Will every font come with a free order of chicken wings?

[Jackson]

Questions:

Does your iTunes model mean you only take 30% commission?

How would this work with modern fonts (i.e.: OT Kerning and OpenType features)?

What about version control, particularly with customers who need to share documents?

Have you done any research that suggests there are enough customers setting only headlines for this model to make any money for a foundry partner?

[hrant]

BTW: https://typography.guru/forums/topic/101389-forwarding

hhp

[oldnick]

Frankly, I don't think I would be persuaded to purchase using this particular marketing models, but I am an old curmudgeon who's hopelessly out of touch with life in a post-PC world—which means, sure: why not? It does appear that many adopters of the later-and-greatest technologies don't always make the soundest financial decisions...

http://www.washingtonpost.com/business/economy/new-ipad-users-slowed-by-...

Plus, if this process offers an additional revenue stream, what's the downside of giving it a chance?

[jasonc]

Ummm, not to throw cold water of tech on your plans, but ... hinting? vertical metrics?

Jason C

[brianskywalker]

Jason, what about them? You include only what is necessary for the purchased glyphs. Why include the hints for 'a' when you only bought 'c' and 'd'?

[JamesT]

So if I were to buy "abcd" I presume I would download a font containing "abcd" if I then return and buy "efghi" would I get a new font containing "abcdefghi" which would require me to uninstall "abcd" and hope the versions don't conflict or that all caches have been cleared (even though that isn't always a guarantee).

How would the average user go about dealing with multiple variations of the same font?

This, to me, is where it is different than music files.

[jasonc]

Jason, what about them? You include only what is necessary for the purchased glyphs. Why include the hints for 'a' when you only bought 'c' and 'd'?

which would be true, if all the hinting information was stored in each glyph. But True Type hints (of any value) reference control values stored in the cvt table, meaning you need to parse the CVT to decide what parts you want to keep? And then adjust later if they buy more letters?

And what if I buy a font with no accented characters, are the vertical metrics set for that "font"'s tallest glyphs? But then I come back later and buy a capital A+ring, does the font's vertical metrics change for the new taller glyph? That would respace my documents created using the font.

[brianskywalker]

Vertical metrics are set font-wide, aren't they? So I think no matter what glyphs you get they'd be the same.

Note: I don't know that I like the idea of this project or not.

[David Jones]

Ryan Maelhorn: so umm, maybe you should tell us why and how this will revolutionize font sales on the web? Will every font come with a free order of chicken wings?

Hey Ryan! Our goal is to make high-quality commercial fonts more accessible to the average consumer. If that means including a free order of chicken wings with every font, then fire up the oven and break out the barbecue sauce! FontSlice will be the first digital type distributor to significantly lower the average transaction cost of a font without devaluing the font itself. We think FontSlice will allow a whole lot more people to enter the wide world of type — and that would be revolutionary.

[Florian Hardwig]

Well, as the skeptics already have spoken (consider me one too), I’ll play the devil’s advocate here and picture a scenario where this kind of service might be of benefit:

As a graphic design studio, we receive a number of logotype files from different institutions/companies, to include them in some brochure or poster. Dealing with graphic laymen, the problem is that these files aren’t fit for print: pixelated GIF, embedded DOC, you name it. Very often, this is about a few characters only, from an easily identifiable and common typeface, without any modifications. It would be a matter of minutes to download the font and re-set the damn line. We don’t do that, because we don’t want to spend money on lame ‘classic’ fonts that we don’t like, and won’t ever need again. I prefer spending money on good fonts.

So, what to do? Lose hours of time trying to reach someone who can send us a good file? No, that’s hopeless (“But I already sent you our logo twice!”, “Just grab it from our website! It’s on the top left.” “Speak to our IT guy. He’s back next week.”). Simply send the crappy files to the printer and not care? Hardly.

This may sound far-fetched, but it happens all the time. In these cases, I wish for a pay-by-the-letter service. As a workaround. The critical points raised (versions, kerning, OT features, hinting) all hardly apply.

I doubt that this is enough to build a business model on. But maybe the target group of FontSlice is not us professionals, anyway?

[daverowland]

Yes as part of my old day job I had to vectorise loads of crappy pixelated logos. If we didn't have the font, it was often way more cost effective to trace the characters needed (it's what got me into making fonts in the first place) than buy the font. In this situation a pay-by-glyph system would be useful. I wonder how big this specific Market is. If this system took off I wonder how long it would be before section, paragraph, currency symbol and all the other virtually pointless glyphs we feel obliged to include in our (display) fonts cease to be made.

[David Jones]

Great questions Jackson. I hope the following helps.

Jackson: Does your iTunes model mean you only take 30% commission?

Our commission can be as low as 30%, but the exact percentage will depend on a number of factors, such as the level of exclusivity (there will be three) and whether or not the font is being actively promoted. We also realize that there's probably not a one-size-fits-all distribution agreement, so we plan on working with each foundry individually to come to an agreement that works for both parties. We want to treat our foundries like royalty, even if that means hiring more employees to handle all of our accounts.

Jackson: How would this work with modern fonts (i.e.: OT Kerning and OpenType features)? What about version control, particularly with customers who need to share documents?

It took months of careful programming to create the algorithm that produces abridged font files. In a nutshell, the algorithm analyzes each line of the original font file and determines what type of information it contains. The algorithm won't even consider altering the line unless it contains spline data, which is the only data within a font file that is absolutely necessary to reproduce the shape of a particular glyph. If a user has rights to a certain glyph, then the spline data is left untouched. If a user doesn't have rights to a certain glyph, then the spline data is removed. All other information (including kerning and OpenType data) is left untouched. As long as the user has rights to the necessary glyphs, all OpenType functionality will remain intact. The abridged font is identical in every way to the original file, save the missing spline data for unlicensed glyphs. This means that version control would be handled the same way it is now.

Jackson: Have you done any research that suggests there are enough customers setting only headlines for this model to make any money for a foundry partner?

It's true that FontSlice will be most useful to users that are typesetting display copy. However, assuming a user is able to predict which character ranges would be required for a particular application, there's no reason why FontSlice wouldn't be used for larger amounts of text. We also believe that users will prefer to consolidate their font property in one place, whether they're buying abridged or unabridged fonts.

[David Jones]

oldnick: Plus, if this process offers an additional revenue stream, what's the downside of giving it a chance?

Nick, you're absolutely right. The intent of FontSlice is not to replace the current model, but rather to supplement it with an additional revenue stream. FontSlice will be the gateway drug for people who might otherwise stick to free fonts.

[David Jones]

jasonc: Ummm, not to throw cold water of tech on your plans, but ... hinting? vertical metrics?

And what if I buy a font with no accented characters, are the vertical metrics set for that "font"'s tallest glyphs? But then I come back later and buy a capital A+ring, does the font's vertical metrics change for the new taller glyph? That would respace my documents created using the font.

Hey Jason. These are very valid concerns and things that we've considered as well. In my response to Jackson's second question, I explained a bit about how the algorithm works. Basically, the information within the font file will remain perfectly intact, except that the spline data for unlicensed glyphs will be removed (metric information will be left untouched). This means that designers will not be required to re-adjust text positioning after upgrading an abridged font.

[David Jones]

brianskywalker: Jason, what about them? You include only what is necessary for the purchased glyphs. Why include the hints for 'a' when you only bought 'c' and 'd'?

Your right Brian. Its really only necessary to include hinting, metrics, and OpenType information for purchased glyphs. The upside of this technique is that it would result in smaller font files. The downside is that it would take a bit more processing power since the algorithm would need to make a few more decisions about what information to include or omit. It's not clear at this point which method is more advantageous on a large-scale platform, and we may choose to alter the algorithm after we gather server statistics for the first few months. In any case, the end result will be the same: a fully-functioning abridged font that is otherwise identical to the original file.

[Nick Shinn]

The “designer fonts” with which you intend to populate your site seem to have been left to the last minute, as if the product, in contradiction, is generic.

In my experience there are two models a new distributor can adopt: either a small, select, carefully curated group of typefaces—or a large enough volume of common styles in which keyword searching can come into play.

The first option grows into the second over time.

For instance, when Apple launched DTP, part of its success was having Helvetica and Times bundled with the LaserWriter software: instant cred.

Recently, Google has launched its web fonts site, with the emphasis on the functionality of the site as a design tool, and the functionality of the font formats and licensing model, but has been criticized for the overall quality of the fonts available. Nonetheless, its fonts are being used; because Google is massive, a benefit you don’t have.

I would suggest you either solicit particular foundries with product that you think will give you a distinct design profile to match and complement your business model (curation option), or approach a foundry/publisher which can provide you with a large collection that spans the major type categories.

Actively seeking specific content would, in my opinion, be better than just saying “open for business” and taking what comes.

[David Jones]

JamesT: So if I were to buy "abcd" I presume I would download a font containing "abcd" if I then return and buy "efghi" would I get a new font containing "abcdefghi" which would require me to uninstall "abcd" and hope the versions don't conflict or that all caches have been cleared (even though that isn't always a guarantee).

How would the average user go about dealing with multiple variations of the same font?

Hey James. Great question. This was actually one of the tests we performed in our testing phase. We wanted to make sure users could easily upgrade an abridged font without too much hassle. To the operating system, two different abridged versions of the same font will appear to be identical. Thus, when you install the upgraded version, it will overwrite the older version. Some software applications will need to be restarted before they will recognize the upgraded font, but this is the case when installing any new font, abridged or not.

[David Jones]

brianskywalker: Vertical metrics are set font-wide, aren't they? So I think no matter what glyphs you get they'd be the same.

Yes, vertical metrics are typically set font-wide. From a coding perspective, its relatively simple to make sure the vertical metrics of an abridged font match the original font file.

[David Jones]

Florian Hardwig: As a graphic design studio, we receive a number of logotype files from different institutions/companies, to include them in some brochure or poster. Dealing with graphic laymen, the problem is that these files aren’t fit for print: pixelated GIF, embedded DOC, you name it. Very often, this is about a few characters only, from an easily identifiable and common typeface, without any modifications. It would be a matter of minutes to download the font and re-set the damn line. We don’t do that, because we don’t want to spend money on lame ‘classic’ fonts that we don’t like, and won’t ever need again. I prefer spending money on good fonts.

So, what to do? Lose hours of time trying to reach someone who can send us a good file? No, that’s hopeless (“But I already sent you our logo twice!”, “Just grab it from our website! It’s on the top left.” “Speak to our IT guy. He’s back next week.”). Simply send the crappy files to the printer and not care? Hardly.

This may sound far-fetched, but it happens all the time. In these cases, I wish for a pay-by-the-letter service. As a workaround. The critical points raised (versions, kerning, OT features, hinting) all hardly apply.

I doubt that this is enough to build a business model on. But maybe the target group of FontSlice is not us professionals, anyway?

daverowland: Yes as part of my old day job I had to vectorise loads of crappy pixelated logos. If we didn't have the font, it was often way more cost effective to trace the characters needed (it's what got me into making fonts in the first place) than buy the font. In this situation a pay-by-glyph system would be useful.

You both bring up a good point. The problem of vectorizing old logos is certainly one of the many applications of FontSlice. In this case, the underlying problem is that your limited usage of an asset doesn't justify the cost. This argument could be made whether you're designing logos or any variety of other media, such as advertisements, blog headers, brochures, covers, invitations, or posters; all of these applications may only require a small subset of glyphs. Of course, if you only purchase 'good' fonts, then you can use them time and again, but fonts are often chosen for specific projects and designers may not want to commit to a font if they're unsure how useful it will be to them in the future. FontSlice lowers that initial barrier to consumption and allows the user to make this commitment over time. While professional designers are certainly a part of the target market, I believe FontSlice will attract people who previously may have never even considered paying for a font.

[aluminum]

I can see the niche use where one needs to deal with a slogan or logotype or sign, perhaps. But I don't quite see how this would have much mass appeal.

So, I think it's a bit silly. But, then again, I thought Typekit was (well, still is) a bit of a silly model as well, so what do I know?

[brianskywalker]

Reminded of Google web fonts: Google web fonts can do the same thing already. It splices up the kerning and hinting data of the fonts and allows you to use any subset of a font, including a single glyph. Of course this is not useful in the same way—all of the Google web fonts are already free. This is only intended to allow you to increase page load times. At the core it uses sfntly.

[David Jones]

Nick Shinn: Actively seeking specific content would, in my opinion, be better than just saying “open for business” and taking what comes.

Nick, the idea of actively seeking specific content is very sound advice. Although it may take a bit longer in the beginning, having the right partners will help legitimize FontSlice and carry it forward. Thanks for your insight.

[David Jones]

brianskywalker: Reminded of Google web fonts: Google web fonts can do the same thing already. It splices up the kerning and hinting data of the fonts and allows you to use any subset of a font, including a single glyph. Of course this is not useful in the same way—all of the Google web fonts are already free. This is only intended to allow you to increase page load times. At the core it uses sfntly.

You're right Brian. The idea of font subsetting certainly isn't new. It's really the business model and platform that is unique. The technology is rather straightforward.