TYPE HomeLogo for TYPE Magazine

Join TYPE

New type

“Good Vanilla”

A conversation with Kent Lew about character sets, multi-script faces, and Clarimo

How do you craft a Latin Extended character set, design all the glyphs to go with it, and finish with a highly legible workhorse typeface? Call Kent Lew. His latest effort, Clarimo, checks all those boxes and more. Managing editor Lucas Czarnecki sat down with Lew to find out how he did it.

Lucas Czarnecki: Roger tells me that you’re the go-to guy for character sets. How did you learn so much about this side of type design?

Kent Lew: I don’t know if I’m the go-to guy, as I suspect each foundry has one. Back when type foundries were starting to think globally and wanting to commercialize in different markets, they all faced the same challenges. One of the main challenges was determining which languages to support and what glyphs were needed to support them. Around Font Bureau and Type Network, I joined in those discussions, along with Cyrus Highsmith and David Jonathan Ross.

Like many foundries, I suspect, we started by looking at the WGL4 that Microsoft established, and then took it from there, eliminating some and maybe adding others. We would evaluate things like the value of a glyph or glyphs, in terms of added language coverage, versus the additional work—whether it was a simple base-plus-accent or a whole new form, including kerning implications. When you’re facing a large family, these things can add up.

An early sketch exploring shapes and proportions.

For me, personally, that process rekindled a latent interest in and an aptitude for languages. Although I am not particularly fluent in any other languages, I have a passing familiarity with French, Italian, Spanish, German, and so on. I even spent some time in my past learning Devanagari script and studying Sanskrit. Turns out, languages fascinate me.

Following the experience with Font Bureau, I continued to do my own research. Over the years, I’ve compiled a number of my own resources, such as an extensive Python library mapping languages to character codepoints. From these, I’ve developed scripts that allow me to, for instance, choose a language and have it select all the necessary characters in the font and report any that are missing. Or select a glyph and report which languages use that character. Or feed in a character set and report which languages are covered.

For Clarimo, had Morisawa gone through that process of discovery already, or is that why they approached you?

I think they’d dipped their toes into the water. When they contacted me, this project was clearly an effort to “up their game” in Latin design and expand into Western markets. But they appeared to lack the confidence internally at that time and were looking for a western designer to partner with and provide guidance. Matthew Carter, who has consulted with Morisawa for many years, recommended me for this project.

Clarimo became many things for Morisawa: It was conceived as the foundational, multi-purpose, highly legible sans-serif in their Latin catalog, both as a standalone family and as the basis for an updated “Latinji” complement within their all-purpose Japanese UD ShinGo family; and it became a vehicle for developing their own extended Latin/Cyrillic/Greek character set.

Regarding the latter, Morisawa already had a vision for the kind of coverage they wanted to achieve, and it was up to me to help them navigate that vision and turn it into a comprehensive character set. There was a fair amount of back-and-forth, where I would educate them about individual characters or present optional ranges to consider, and we would evaluate the relative benefits. Much like what I described earlier. But it was particularly interesting to approach this from the perspective of a foundry that is not coming from North America or Europe. The markets that they might have interest or opportunity in differ in subtle ways.

“For me, personally, that process rekindled a latent interest in and an aptitude for languages.”

In particular, this project gave me the opportunity to expand my knowledge of the extended Cyrillic range, for non-Slavic Cyrillic-using languages, such as those from the Caucasus and Volga Upland or Central Asia, for instance. (Funny side note: Right in the middle of the project, the government of Kazakhstan announced that it would be switching from Cyrillic to Latin orthography!)

The Clarimo development, then, formed the basis of a standard character set and will provide a reference and something of an exemplar for future Morisawa designs. I know that my work on Clarimo informed the process of Matthew and his fellow designers working on the Role superfamily, which was going on simultaneously.

In fact, Matthew would call me in to consult with them occasionally. In turn, questions that would arise from those designers would sharpen my work.

So my engagement had really several parts: This first part of figuring out the character set; then a second phase actually designing the extended Latin, Cyrillic, and Greek for this basic, multi-purpose matching sans; plus some coaching with their internal designers.

As the basis for the Latinji to accompany their UD ShinGo Japanese fonts, the Clarimo project became an opportunity for the internal designers working on that adaptation to become more familiar and confident with Latin, Cyrillic, and Greek shapes.

Tell me about designing a “default” Latin typeface.

Well, I think it takes a certain temperament to design a “boring” typeface. [laughs] But, seriously, I enjoy this type of work. Certainly, Clarimo is not breaking any new ground stylistically. But there are other factors to justify its existence: Compatibility with the Japanese, the wide range of weights and widths, and its role as the foundation for a multi-script superfamily. (Morisawa had other outside designers develop corresponding Arabic, Devanagari, and Thai fonts, building on the design language of the Latin and the Japanese.)

With this kind of brief, my goal was to create a design that would distill the letters to a kind of essential, “Platonic” form. I wanted them to feel completely familiar and natural, like “of course that’s what they would look like,” as if they’d always been that way. The challenge with such a goal is not to end up with something too boring, with no flavor whatsoever. Yes, it’s meant to be vanilla, but hopefully it’s a really good vanilla!

Testing page color side-by-side with Japanese setting.

You were “matching” the Latin to their Japanese. How is this different from your normal work?

Yeah, usually I have worked from more of a blank slate, with either a very general brief from a client or just pursuing my own ideas from scratch. But in this case there was a very specific task of harmonizing with the existing Japanese. Also, with this project, the final production, naming, and marketing fell completely to the Morisawa folks, whereas I might ordinarily take more responsibility for some of this myself or be more involved. It was actually kind of a refreshing change.

For example, the name “Clarimo” was all their doing. I mean, they would run ideas past me and we would bounce possibilities around. But ultimately, they had to finalize this through their own process and with their own trademark searches, etc. I like where they wound up: “Clari” for clarity, and the “mo” suffix alluding to Morisawa.

In terms of matching with the Japanese, the Morisawa folks gave me a brief setting out some of the stylistic aspects they envisioned. They characterized their UD ShinGo as very “modern(-istic),” “neutral,” and “geometrically stylized.” At first, this latter description puzzled me a little.

Clarimo’s full name is “Clarimo UD PE,” with the PE standing for Pan-European. Did we mention it includes Vietnamese?

Looking at the Japanese with my Western eyes, it certainly doesn’t look particularly “geometric” to me. But as I studied it and looked at other Japanese styles, I could start to understand how from within a culture so much more steeped in a strong calligraphic tradition, even in their types, the simplified, constructed and rationalized forms of the UD ShinGo could seem geometric. Nevertheless, I felt that a Western interpretation of geometric for the Latin would feel too severe in the intended context. So, my approach focused on simplicity, balance, and a certain symmetry that can accompany a geometric style, but without taking it too literally.

“My approach focused on simplicity, balance, and symmetry—without taking it too literally.”

Perhaps the more difficult challenges of creating a Latin face to accompany a Japanese were the technical ones. A typical Latin font will usually look small next to a Japanese set at the same nominal size. To make a Latin that would pair reasonably with Japanese meant designing fairly large on the body. But the vertical proportions that Morisawa initially proposed were quite large. Matthew and I both felt that this would fall outside any acceptable norm for a standalone Latin font. So, there was a period of exploration to find just the right proportions to balance needs with expectations.

This was also a factor in matching texture and color. The Japanese script has an inherently wide range of glyph density, from the loose and open hiragana to the sometimes quite crowded and dark kanji. So a Japanese text setting has a very different texture that can look spotty to Western eyes, and it can be challenging to determine what is an equivalent Latin setting.

The development of the condensed widths to match the Japanese system posed another new challenge. Japanese characters are typically designed to fill a constant, square space. So scaling a Japanese design is an almost mathematical process. The UD ShinGo family has condensed fonts that are 90%, 80%, 70%, 60%, and 50% width, in which the character space is literally a rectangle of 90%, 80%, etc. width.

Although there is some stem and stroke compensation, it seems to be less of a concern, and condensed heavy styles can even start to show a reverse stress.

Latin doesn’t scale so linearly. How do you make a 50% lowercase i or l? An m or w narrowed to 50% quickly goes out of acceptable relationship with a 50% n or o, in terms of both proportion and density. And it’s not simply a matter of compensating stem weights.

So this development involved a lot of testing to find the right parameters.

What additional challenges did you face while designing the Cyrillic and Greek, considering you do not read any language which uses those scripts?

Any type designer who ventures beyond their native script faces this challenge. I think we all grow up to form an inherent sense of what words in our own language look and feel like. Type designers become particularly attuned to this, and we’re constantly evaluating proportions and spaces against this innate sense of rightness.

But with a language or script that is not our own, we don’t have that same internalized reference. So, I try to compensate for that by immersing myself in examples of various languages and scripts. I’ve collected some books in different languages; and even if I can’t read the language, I’ll spend time just absorbing the unique shapes and trying to internalize the rhythms of the language itself.

“These qualities lend a different sense of space, rhythm, and color to each script, a distinctive character to their words.”

Another resource that I’ve developed over the years is a collection of extensive word lists for various languages. From these I can use Python scripting to compile test files with real-world examples of certain combinations, or even a massive file with example words for every possible pair (that one took hours to run!). I always want to proof with real words, rather than abstract strings, to let the whole inform the parts.

Of course, I also relied upon review and valuable advice from colleagues with expertise in these scripts—Maxim Zhukov, in the case of Cyrillic, and Gerry Leonidas, for Greek. They helped provide that innate sense of proportion and rightness, which was paramount for this family.

The official Clarimo UD PE specimen includes its sister faces: Clarimo UD Arabic, Clarimo UD Devanagari, Clarimo UD Thai / ThaiModern, and others.

It’s important to become attuned to the unique character of a different script, learning what kinds of space are inherent in each. So, for instance, with Cyrillic, noticing that there’s a whole lot more activity going on in the interior region between the baseline and the x-height to distinguish glyphs.

Latin has considerably less going on there, with much more activity above and below in the arches and extenders. And Greek is quite unlike either, with its diverse range of more organic, dancing shapes.

All these qualities lend a different sense of space, rhythm, and color to each script—a distinctive character to their words. And all these differences have to be respected and accommodated. I hope to have achieved that with Clarimo.

Previous entry: On a role