Typeface Design Evaluation Criteria

The quality assessment (QA) or the evaluation of the quality of a typeface design is a very complex task, as it requires assessing the design’s technical, contextual, and aesthetic qualities.

Gerard Unger (2018) mentions four main qualities:

  1. Functionality (the type’s technical features);
  2. Convention, or social quality (the contextual application);
  3. Creativity (originality and Aesthetic quality);
  4. Expressiveness (display or emotional design);

We will concentrate on the first three, as they seem to be the most important and assessable ones that seem to correspond to the main criteria used (more on this near the end of the article).

Some qualities are more objective or straightforward to assess due to their technical or historical nature. Qualities such as functionality or originality. Others, such as the correspondence to its social context of the (reading) application or artistic expressiveness, are more subjective. This is due mainly due to the human or artistic nature of Design.

1. Functionality

Especially in our curricular context — the design of a Text Typeface Revival — some of the type’s features are easy to see and to assess.

1a) [Color] Consistency :

Does it render well? In our case — for printed text typefaces —, does it print well on 8-12 pt text sizes? This means that it should produce an even gray color/pattern on the page.

Irregular spaces or ink blots are a sign of lack of quality. These are usually due to incoherence of the design system parameters. E.g. irregular stem and bowl stroke widths, different proportions of the shapes, inconsistent use of measurements, etc. Or to the inadequate technical design of the shapes. E.g. poor point placement in vector drawing, uneven or improper curve tension (e.g. in the shoulders or bowls), lack of visual compensations such as tapering, and overshoots, etc. Or even to poor handling of black and white ink ratio. E.g. inconsistent [basic] spacing of the glyphs, or the lack of proper kerning when necessary. These are mainly technical drawing aspects that should be easily fixed.

[update: 2022-08-04]

Today, during a calligraphy class, a colleague — PG — asked our calligraphy teacher if there is any sort of checklist or quality evaluation guide to check against our [calligraphy] designs.

The answer was… it’s tricky. But the first hint was the family likeness of the curves (as in Edward Johnston’s approach)

[future update: a post on Johnston’s 7 rules for consistent calligraphy]

So, here’s a first draft tackling this issue: a visual/technical checklist:

  • Vertical Metrics: check your measurement lines
    • x-height,
    • cap height,
    • ascender height, and
    • descender height lines

      Check for coherence and adherence to the desired design/model);
  • Stroke width & stress (or pen nib angle): again, check to see if the stroke width is coherent throughout. This means that
    • bowls should be the widest,
    • verticals the “normal”,
    • diagonals thinner, and
    • horizontal strokes the thinnest.

      How much depends on the design model, as it will imply a shift in the pen nib angle that produces the [shadow] stress of the shape.

      Eg. the roman letter form implies a 30º pen nib angle (with slight variations). It results in a ~2:1 proportion between verticals and horizontals. And the stress of the letter will be also around +30º and -30º for the fattest and thinnest parts of the round letter shapes. There will be slight nuances, such as wasting on the letter stems and slight build-ups on the bowls, but these are to be taken into account later on in the optical corrections ;
  • Curves & Arches: as our calligraphy teacher — Christopher Hannes — mentions, always check the family [likeness] coherence of the arches.
    • Rounded
    • Pointed or broken (as in fraktur?)
    • Semi-pointed (hybrids and bastards)
    • Others (uncials, Legende,…)

      Don’t mix’n-match different arch designs
  • Direction of strokes. And also entry, connecting, or terminal strokes. Keep in mind, that writing, calligraphy, and eventually typography all inherit qualities from the writing tools. So, sometimes, it is important to know writing/calligraphy models to understand their shapes. Coming back to the checklist, check for coherence (within the design) and historical/technical adherence of
    • entry strokes (are they facing the same logical [hand/writing instrument] ductus direction?
    • angle or inclination

    • The main issue that arises ALL THE TIME is the direction of serifs and entry strokes — namely the “d”. Or the tension of the “b” “p” “d” “q” bowls… It shouldn’t be geometrically rotated or reflected. It should follow the writing direction. Please. Check. Them. Strokes. [image here please!]
  • Horizontal metrics & proportions: This is the most difficult to explain nowadays. Counter shapes should be:
    • Regular, or
    • Rhythmically [inter] spaced

      Currently, there is a widely accepted approach to make all the widths “feel” similar [reference please!]. This means that all the strokes and [white] counter shapes of the letters are regularly interleaved. This is the easiest way to produce an “even” gray color on the page. But this wasn’t always the case. And it shouldn’t. As Frank Blokland points out [need a reference here!], sometimes, changes in the horizontal rhythm are also beneficial for the reader. The trick is to get them working as a [coherent] system. This was more common in the classical (as in Old Style / Venetian) periods. Roman Caps also follow a very varied proportion system, with the letters’ set width ranging from 2:2.2 to 2:1 proportions. So, whatever your approach is, make sure to set a design model and stick to your proportions. Coming back to the “cheat guide” mode, a simple and modern way to address this is to think that when you decide on an arch form — that’s why the “n” is such an important letter — it will create a specific counter width that will set the base horizontal rhythm to follow. Meaning that, if you design a humanistic [foundational] hand n, the arch will be of “roman” regular/squared/circular proportions. When you analyze your design, it’ll mean that the width of the inner shape —the counter— of the n, will be from 2,5 to 3 stroke widths (also known as pen [nib] widths, or pw). This amount of “air” or white space will have to be roughly the same on all counters [for a modern, regularly interleaved design]. That is why the counter of an “h” is roughly the same as the “n” [see the visual compensations]. But the “o” has to be wider (it has an extra arch on the bottom that “robs” the air out of the counter and needs to be compensated)
  • Spacing: this one is kind of easy… the spaces around the letter, the sidebearings, should be roughly from 1/4 to 1/2 of the width of the counter of the letter. This is using Tracy/Karow/Cheng’s rules of thumb. It just means that the spacing should be proportional to the “color” of the letter. In other words, the side bearings reflect the ink-to-paper ratio produced by the proportions of the stem width and the rhythm of the counters of the arches. Tighter arches = smaller counters = thinner side bearings. But it is not always easy. On very light [stem width] designs this has to be “almost reversed” [more on this sooner];
  • “Closeness” of the apertures of the shapes:
    • open, or

      This one is kind of related to the curve tension and spacing. It means that it could be considered a product of mixing the two. It just means that when you decide on a design or a historical model, you have to decide to coherently keep shapes more “open” or more “closed”. Just picture a “c” you can have a fairly open classical Monumental Roman “C”, or open sans serif humanist Frutiger “c”. Or you can have a “closed jaws” sans serif grotesque Impact or Ionic “c”. Closing these curved apertures also means closing the arches of the “a”, the “e” the “2”, the “g” etc…
  • Joints & cross-strokes: Thick with thin strokes always, please. Maybe thin with thin. But avoid crossing thick strokes. If needed, just move them close to the point of [almost] touching, or producing an evident corner joint [image needed here!] Avoid crossing or touching strokes too much.
  • Optical compensations: Beyond metrics and stress,
    • when two strokes meet, black/ink overload should be avoided (strokes must be thinned, or tapered).
    • Lookout for illusions, such as the crossing directions of the X — the thin one isn’t really crossing the thick one… it’s more of a two-stroke thing—, asymmetry of serifs (width and thickness), apparent alignment of smooth curves, vertex corners, and straight edges. As a rule of thumb [see Peter Karow’s breakdown of these rules either in his book or in Karen Cheng’s manual] squared shapes should be within the measurements (eg. x-height), round shapes should be visually compensated (eg. from 0.4 to 1.2%) beyond the measurement lines (stick coherently to your decisions!) and [triangular, acute or sharp] corners should go slightly beyond that.
    • And finally, evaluate the “amount” of ink a specific letter has. An “n” has the same strokes as an “h” right? Almost… the ascender means that it also has extra ink to deal with… so spacing —namely the side bearings— has to be adjusted accordingly.
  • [more on this checklist soon]
[insert here a checklist based on the links below and the 1st assignment syllabus]

There are other less-easy-to-spot design issues such as inappropriate [Latin] proportions, the lack of adherence to historical standard shapes — in our case, an easy thing to spot as we are designing text revivals —, or inconsistency of the design model chosen (direction of serifs, types of terminals, and apertures, etc.) These are easily “fixable” by studying historical, well-known models and the original calligraphic hands that inform the shapes.

Finally, too much expressiveness may produce irregular text results. But this one is not so objective to assess…

1b) Purpose appropriateness:

Does it respond to the technical or to the design briefing requirements? E.g. designing a typeface that works well on an encyclopedia or dictionary (that usually requires variations of slope, weight, and caps) without sacrificing space or the final visual quality (sometimes the design has to be adapted to fit a particular size, or to be reproduced in a particular paper or screen conditions).

1c) Legibility:

Does it “acquire” well at a 40-50 cm reading distance? This means all characters should have robust and unequivocal designs. It’s the “Ili1” and “ocea” likeness test. But applied to the whole character set. Consistency in the design, but uniqueness and/or variety in the glyph’s shapes.

1d) Character set (completeness):

For any (text) typeface, a complete character set is required for the language it is supposed to work in. Usually, this means developing (at least for us) an Opentype Std Caracter set. Rule of thumb, test it with real content. Majuscules, minuscule, diacritics, numerals (figures) and symbols. In 2022, we’ve used Monotype’s minimum recommended character set.

Completeness also means that a typeface should be a [part of a] family. This means it should have consistent stylistic siblings such as weight, width, and optical and italic fonts (or masters) whenever possible.

Not required, but highly desirable is the presence of appropriate and relevant Opentype Features (Contextual, Stylistic and proper additional features).

Also desirable is the capability of performing well on screen and on print. This means having decent hinting.

How to access the type/font’s proper functionality? As Typetogether have put it, you must test it (some foundries provide test or rental versions of their fonts)

2. Convention (Social quality)

“The social quality is largely due to convention, …” mentions Unger (2018). In this case, is it something the reader is expected to see or read in this context? We know from previous legibility research (reference needed here) that it takes as little experience as a 20-minute period for users to display similar results with known or comfortable typefaces. In any case, designing for a specific group of users (children, vision-impaired), or contexts (brands, business, human endeavor areas such as music or other) means that, although we adjust easily to new or different designs, we also need to address and manage expectations and the typeface’s visual qualities.

[insert The Crystal Goblet argument here]

2a) Familiarity

Hence, we must know and make use of familiar designs based on historical conventions. E.g. the use of Roman Capitals for Hollywood blockbusters. Or the use of blackletter for highly institutionalized academic documents. These are just a couple of examples that take convention into account. But one can/may/should (?) break conventions to introduce surprise or (branding) uniqueness.

Whatever the approach, always take into account that social and historical convention(s) are often better received/perceived due to their “degree of recognizability and familiarity”.

2b) Contemporaneity

One additional note to take into account is the contemporaneity of the Design. As with any given design project, it should be the product of our time and culture — even for type revivals as we are doing in this course. This means, that even when reviving something from the early 20th century, we should infuse it with the features, requirements, or “feel” of our own cultural context and time. This is, of course, one of the most subjective aspects to be addressed here. Nonetheless, one of the most important. Case being the difference between facsimiles and revivals. Let’s take The Fell Types that have been (brilliantly) digitized by Igino Marini. These were scanned and then chosen the best glyph representations were included in the digital typeface. But it consists of facsimiles. They’ve kept all the vicissitudes of the original print. But, if you take Jonathan Hoefler’s Requiem, it takes the “original XVI century design concept” into a contemporary digital [leaner, brighter, cleaner] typeface.

As such, for the convention or social quality, we should consider the context (when, where, and to whom) it is being designed.

3. Aesthetic quality (creativity and originality)

This quality, although it might seem difficult at first, is a fairly simple and straightforward aspect to assess:

3a) Context-based originality

In the first approach, one has to determine the context of the operation of the typeface. Which (social) group of users will see it? Is it a familiar face to them or not? This is important because its shapes might be unfamiliar. E.g. Merovingian to non-french or wild style to non-graffiti connoisseurs might seem totally novel to some. But not to the people who are used to seeing and writing them.

The introduction of these kinds of visual and/or historical features has to take into account the (social) convention expected.

3b) Design-based creativity

Here lies the true creativity or originality. There have been few moments in history where true innovation occurred.

[examples needed here]

Yet, many original or innovative typefaces have been developed, usually combining, iterating or modifying features from their predecessors.

Hence, in this item, we have to take into account Margaret Boden’s two main types of creativity: Personal or Psychological Creativity (p-creativity); and Historical Creativity (H-creativity).

[reference needed here]

The first is the type of solution that is novel or innovative in a given personal, authorial, or group context (such as in the learning context). We might come up with solutions that have been tried and tested before, but we need to develop them for our own propaedeutic needs or process. These may be surprising and valuable solutions for the author or persons that come up with them. But add very little to the overall context of the discipline or overall state-of-the-art knowledge.

The second one, and one more difficult to achieve, especially in time and resource-limited contexts such as academia, is the “true” creativity or originality. It signifies that the designer must come up with a truly innovative solution — whether by combining, modifying, synthesizing, or creating a new design —, that hasn’t been seen in the history of the discipline. Again. There have been just a few moments of true innovation in [Latin] writing and type design.

As Unger says, originality may clash with convention. As he cites, it may be interpreted as a “sophisticated bending of the rules or conventions”.

In the learning process, copying, modifying, and imitating are acceptable actions — when properly credited and consciously identified—, as they actively contribute to the improvement of our skills and promotion of culture as a whole. But “homage”, “pastiche” and “historical or visual reference” are very blurry areas of operation. And ethically tricky ones (not to mention often illegal) when done professionally.


These three qualities consist of the main criteria to evaluate a good typeface.

It is worth mentioning a final note on Unger’s fourth quality: Expressiveness (display or emotional design). This final quality is expressed by Unger as the “interpretable concepts” that fit its purpose or social context. As such, it seems to be a combination of previous objective and subjective factors. Although, as he mentions it, and it is no doubt an important one, in our view, this should be one of the results of the evaluation and not an evaluation criterion by itself.

One other thing we are not covering here is the Font [as in digital software] quality assessment. Proper family and style names/linking, consistent metadata (copyrights, blue values, WinAscent, and other important technical meta-information within the font); adequate font formats (OTF, Woff, OTVar,…) are also very important aspects, but they should be addressed on a different occasion.

[insert font production/engineering Q&A link here]

Also, other authors such as Indra Kupferschmid mention the price of the typeface/font. This is not an issue in our context but should be taken into account when choosing (and trusting) a type supplier. The better or the more recognized the designer, the higher it is expected the price to be… But I feel this is an “everyday” issue that can be applied to buying or renting software, online courses, clothes, or hardware. And is not type-specific… Use your common sense!

Also, with the price, comes the support. Usually, smaller independent type foundries provide faster and better support to address particular issues or client requirements. But, to this day, having bought from Fontshop, MyFonts, R-Typography, and rented from Fontstand I haven’t had any issues yet.

In the end, assessing the quality of the design of a typeface just depends on your own [type] design education. As with everything in life, knowledge and education makes a difference and allow you to make better decisions. And we hope this course is a part of it!


Additional notes and references on assessing a typeface quality:

[insert the bibliographic references here]

By Pedro Amado

Pedro Amado. Assistant Professor at FBAUP. Integrated researcher of the i2ADS research institute.