Flawed Typefaces

What constitutes a flawed typeface? For this article it is defined as a typeface that is perfectly fine—except for one nagging aspect, usually a single character. A flawed typeface is one that either you avoid using entirely because of this lone defect; or one that you use sparingly—and only then, after some alteration of either your design or the face itself to ameliorate the “flaw”. Flawed typefaces are not bad or even mediocre. The whole premise here is that they are good, perhaps even classic or wildly popular. And yet there is a single character that ruins them or, at the very least, causes one to pause before specing them. This article attempts to explain the flaws in 23 fonts: what they are, why they matter, and what to do about them. Disclaimer: these are opinions, not facts. Ultimately, flaws are in the eye of the beholder.

Goudy Oldstyle (Frederic W. Goudy, American Type Founders, 1916)
American Type Founders altered Frederic W. Goudy’s design to fit its common line, a measurement instituted to insure that all of its typefaces could be aligned with each other. Most of the descenders survived the abuse, but not the g. In the digital world there is no reason why Goudy Oldstyle cannot be restored to its proper look.

Goudy Oldstyle compared to Adobe Garamond Pro.

Goudy Oldstyle detail from 1923 ATF Type Specimen.

Bembo (Monotype Corporation, 1929; based on the 1495 roman type of Francesco Griffo)
The hot metal version of Bembo came with two versions of R, one with a short leg intended for text composition and the other with an elegantly extended leg intended for titling purposes. Unfortunately, in the phototype era the second R was the only one available and that situation continued with the PostScript version of Bembo. Both Rs are once again available with Bembo Book Pro. Unfortunately, the long-legged R remains the default character.

Top line: Bembo, Bembo Book MT Pro and Bembo Book MT Pro with alternate R; Bottom line: FF Quadraat and FF Quadraat kerned

Detail of De Aetna (1495) typeface by Francesco Griffo. The model for Bembo.

FF Quadraat (Fred Smeijers, FontShop, 1992)
FF Quadraat is another font with a problematic R. Its leg sticks out a bit less than Bembo’s but it also dips noticeably below the baseline. It spaces poorly and creates a visible tic in a block of text.

Centaur (Bruce Rogers, private 1914; Monotype Corporation, 1929; based on the type of Nicolas Jenson, 1470)
The R in Centaur also has a leg that juts out, but the flawed letter is the odd j with its dirk-like wiggle. This letter is a Bruce Rogers invention since Nicolas Jenson did not have a j in his typeface. Presumably, the wiggle was added to keep the j from having to be kerned. It is distracting in text, beguiling in headings and logos (see john varvatos).

Centaur MT Std compared to Adobe Garamond Pro; both kerned.

DTL Dorian (Elmo van Slingerland, Dutch Type Library, 1996)
DTL Dorian, a font that I recently included in my column on underappreciated fonts, is afflicted with a bothersome S and s. The top curve ends in a bracketed serif while the bottom curves simply curls up, giving both letters a top-heavy look. They don’t fit in easily with other round letters, either.

DTL Dorian T compared to Adobe Garamond Pro; both kerned.

Optimo Didot the Elder (François Rappo, Optimo, 2004; based on an 1819 type cut by Vibert for Pierre Didot)
Didot the Elder is full of weird—yet historically accurate—characters: C, G, S and s with arrow serifs; f with a “serif” on its crossbar; y with a bent descender; and g with a seriffed open loop. Most of these are fun. But the g is annoying in a lengthy text. Unfortunately, there is no normal alternate available.

Optimo Didot the Elder compared to Linotype Didot.

detail of page from 1819 type specimen by Pierre Didot; type cut by Vibert

Sabon (Jan Tschichold, Stempel, Linotype and Monotype, 1967; based on the types of Claude Garamont) and Sabon Next (Jean François Porchez, Linotype Library, 2003)
Sabon, one of the most beloved typefaces of the 20th century, was originally designed for three technologies: foundry, Linotype and Monotype. When it was adapted to film and subesequently to digital it was the Linotype design that was carried forward. This meant that the italic was compromised since the linotype version was duplexed. Thus the italic is wider than normal. This dismays many people, but it might be acceptable if it were not for the nearly round o. The solution is to use Sabon Next Italic instead.


However, Sabon Next is not beloved by fans of Sabon. Not only did Porchez go back to the foundry version of Sabon, but he went beyond it to the typefaces of Garamond and Le Bé which Tschichold was using as models. And then he went even further and added alternate characters (some reprising those of the original Sabon, others more fanciful), swash characters and quaint ligatures. To many, the result is a typeface that seems more Adobe Garamond than Sabon. But, on its own terms Sabon Next is a fine face—except for the alternate q. This is a capital form (common in Renaissance calligraphy) that looks totally out of place, especially with its too short tail that looks shriveled up from embarrassment. But it’s an alternate character and can be easily ignored.

Top line: Sabon Next LT with alternate q and Sabon Next LT; Bottom line: Sabon Italic compared to Sabon Next LT Italic all kerned

CC Galliard (Matthew Carter, Carter & Cone Type, 1992) and ITC Galliard (Matthew Carter, International Typeface Corporation, 1978; based on the types of Robert Granjon)
As admired as ITC Galliard has been from the outset, its aggressive italic has not always found favor, especially the pelican-jawed g. Anticipating this, Matthew Carter designed an alternate g for such people, but when the International Typeface Corporation took the font over from Mergenthaler Linotype, it was set aside, along with other additional characters. These characters were all returned to the font when Carter re-released Galliard (the regular weight only) through Carter & Cone Type. I am not one of those who disliked the pelican-jawed g, but I am happy to have the alternate g as well.

Top line: CC Galliard italic compared to CC Galliard Italic with alternate g; Bottom line: Adobe Garamond Italic for comparison all kerned

Comenius Italic (Hermann Zapf, Berthold GmbH, 1976)
In both kurrentschrift and English roundhand the w is often made in the form of an n joined to a u. This form is usually dropped from modern typefaces—compare Matthew Carter’s Snell Roundhand to the sample alphabet by Charles Snell—but an exception is Comenius Italic by Hermann Zapf. Although Zapf has redesigned the character at least once since the typeface’s debut, the w still looks peculiar to modern eyes.

Comenius Italic (kerned) compared to Palatino Italic.

Roundhand alphabet by Charles Snell from The Standard-Rules of the Round-Hand and Round-Text Hands (1714).

Weiss Antiqua Italic (E.R. Weiss, Bauer, 1926)
Neither Latin nor Italian use y as a letter. Thus, it does not have deep historical roots. In cursive writing it often takes two forms, one derived from v and the other from u (letters that were interchangeable in Ancient Rome). Italic typefaces often follow the v form, often to the detriment of the rhythm of the letters and thus to consistency of spacing. Perhaps the most egregious example of this is Weiss Antiqua Italic (originally Weiss Kursiv), one of the first typefaces to follow the chancery model. The y has a beautiful, sweeping left arm which makes proper spacing difficult and, in the case of ry or ty, nigh impossible.

Top line: Weiss Italic (tracked) and Weiss Italic (tracked and kerned); Bottom line: Cataneo Light Italic for comparison (kerned).

Franklin Gothic (Morris Fuller Benton, American Type Founders, 1902) and ITC Franklin Gothic (Victor Caruso, International Typeface Corporation, 1980)
Franklin Gothic is one of the most iconic of American typefaces. It is an unusual design in that it is heavy and there was never a regular version from ATF. Instead, ATF made condensed and wide versions and then lumped Franklin Gothic with several other Morris Fuller Benton-designed gothics (Lightline Gothic, News Gothic, Monotone Gothic and Alternate Gothic) as a de facto family. This satisfied printers and designers for decades. But when ITC licensed the face they decided to not only adapt it for photocomposition but to make a proper family. Thus, we have the anomaly of ITC Franklin Gothic Book. The updating of Franklin Gothic involved a number of small annoying adjustments that chipped away at its identity, most of which can be accepted as the price paid for having a harmonious family. The one change that went too far was the mucking about with the  distinctive Franklin Gothic g. In ITC Franklin Gothic it looks like a mailbox flag. This completely changes the appearance of a block of text as the ear on the g keeps popping up like a schoolchild overly eager to answer a question.

Top line: Franklin Gothic No. 2 Roman compared to ITC Franklin Gothic Heavy bottom line: Ellington for comparison all kerned.

Thesis (Luc(as) de Groot, FontShop, 1994 on)
Luc(as) de Groot’s Thesis is perhaps the largest type family ever created (144 fonts at last count). Despite its widespread popularity it has one character that makes it unsuitable for certain usages, such as information design: the Q. It has a detached tail which, at best, is a distraction and, at worst, makes the letter look like an O with an accent. This is especially true at small sizes.

top line: TheSans and TheSerif bottom line: Frutiger 55 all kerned.

Syntax (Hans Eduard Meier, Stempel, 1969) and Linotype Syntax (Hans Eduard Meier, Linotype Library, 1997) (now called Syntax Next)
Revising a classic typeface is a dangerous thing, even when there is widespread agreement that the existing version needs a sprucing up. It is even more traumatic to type users when no one but when the original designer thinks that a total overhaul is required. One such instance is Syntax which Hans Eduard Meier, its designer, refashioned from scratch in 1997 as Linotype Syntax. Rather than just fix the poor digitization of the original Meier took the opportunity to return to his original 1955 vision of what the face should be. The result is a design that is more calligraphic than typographic—nice, but flawed. Two of Syntax’s hallmark letters have been ruined rather than improved. In Linotype Syntax the splayed M is splayed even more than before and the R now has an open bowl.

Syntax compared to Linotype Syntax.

Carter Sans (Matthew Carter, Monotype Imaging, 2010)
Sometimes a flaw in a typeface is a very tiny thing. With the new Carter Sans the square dots on the i and j appear too large. Matthew Carter may have done this deliberately to make them hold up at small text sizes. But at display sizes they are too noticeable.

Carter Sans Regular compared to Optima Nova Medium.

Bickham Script (Richard Lipton, Adobe, 1997; OpenType version 2004)
The PostScript version of Bickham Script had a significant flaw: the T, the second most common letter in the English language. It is designed with a very wide base that makes it ambiguous. Is it an T? an I? or even a Z? Instead of being corrected in the OpenType version, this problem has been exacerbated. There are no simpler alternate forms, only two fancier Ts, both of which have similar ambiguity. The designer of the logo for The Astor, an apartment building on the Upper West Side of Manhattan, took matters into his own (or her own) hands and simply lopped off the base of the T, among other indignities.


top line: Bickham Script with regular T and contextual Th; Bickham Script with medium T and different h; Bickham Script with large T and different h; Novia Script bottom line: Bickham Script with contextual swash Th; Adobe Caslon Pro with contextual Th; Big Caslon.

This digression was occasioned by the appearance of the automatic Th ligature in Bickham Script Pro when setting samples for this article. This is a contextual ligature, one that is word-dependent. It is on by default but can be turned off. Adobe has pioneered the Th ligature and the first time that I encountered it I was thrilled. It is exactly what I have always done as a calligrapher to solve the horrible gap created by h following T—one of the most common occurrences in the English language. But, when I began to see it in non-calligraphic fonts such as Adobe Caslon I found it distracting. We have become so inured to that large space that no longer notice it, especially given that we read in chunks and the most common Th words (The, This, That, There, These, etc.) are thus gobbled up easily. Ultimately, a ligature is intended to improve spacing so that the reading experience is not disrupted. A ligature that makes us stop and take notice is not doing its job. Maybe the Th ligature will become second nature, but for now I am having second thoughts about its value.

Burgundica (Gerrit Noordzij, The Enschedé Font Foundry, 2009)
Burgundica is a modernized fraktur. As such, it has several letters that are inherently problematic for modern readers, especially non-German ones. The A, D and S are all extremely difficult to recognize in frakturs. For Burgundica, Gerrit Noordzij simplified each of them, but not enough. Which one is the fatal flaw depends on the reader. For those who want to still use Burgundica, the solution is to find a compatible roman face and substitute its capitals.

top line: Burgundica X025 bottom line: Burgundica X025 with 84 pt Milo Serif Medium S in second example (kerned).

Wilhelm Klingspor Gotisch [Wilhelm-Klingsporschrift] (Rudolf Koch, Klingspor, 1926)
Texturas are not as difficult to use as frakturs for non-German speakers, but they still pose problems. Roman lowercase letters with diagonals in them either stick out or are adapted so well that they are not easily identifiable. The digital version of Wilhelm-Klingsporschrift (now called Wilhelm Klingspor Gotisch) by Linotype was modernized to make Rudolf Koch’s masterpiece more palatable to modern designers. The k is excellent but the x is awful. Vertical rhythmn is essential in a textura and the x is a stumbling block. And, for some reason, Linotype ignored the unfamiliar A. For those used to blackletter, the x is the flaw in Wilhelm Klingspor Gotisch; for those not used to blackletter, it is the A.

top line: Wilhelm Klingspor Gotisch bottom line: Flamande for comparison.

complete character set of Wilhelm Klingsporschrift (Rudolf Koch, Klingspor, 1926). Metal setting courtesy of Michael Babcock.

Univers (Adrian Frutiger, Deberny & Peignot, 1957)
Sometimes the character flaw in a typeface is less obvious. Instead of a letter, it is a figure or a punctuation mark. Univers purists like Helmut Schmid and Willi Kunz insist on Berthold BQ Univers among digital options. They disdain other cuts for being poorly digitized, for having italics with the wrong slope, and for not having the correct ampersand. However, the proper ampersand—a true et ligature—looks odd to those without paleographic or calligraphic training. So, which is the flaw?

top line: Berthold BQ Univers 55 bottom line: Univers SB Roman both kerned.

Gill Sans (Eric Gill, Monotype, 1928)
Figures are often overlooked in discussions of a font’s attributes, but for anyone working in the area of information design or signage they are crucial. The most important figure of all is the 1 because it is the most common. Unfortunately, it is the figure that causes the most trouble. It can be confused with capital I and lowercase l in sans serif typefaces, and in all typefaces its inherent narrow width forces designers to make compromises either in its form or its set width. Eric Gill designed the 1 in Gill Sans without a flag, thus making it virtually indistinguishable from I or l. Monotype resolved this disaster in the days of machine composition with an alternate flagged 1. But it was not included in the digital version of Gill Sans. One solution to this problem was to substitute a 1 from another sans serif face. This is no longer necessary as the flagged 1 is part of Gill Sans Pro.

top line: Gill Sans and Gill Sans kerned with 70 pt Futura Book 1 bottom line: Gill Sans Pro with alternate 1 and Gill Sans Pro with alternate 1 kerned.

Franklin Gothic (Morris Fuller Benton, American Type Founders, 1902) and ITC Franklin Gothic (Victor Caruso, International Typeface Corporation, 1980)
The narrowness of 1 has forced designers to add unwanted space on either side so that it will set properly in tabular matter. This makes for loose text setting. This problem is exacerbated with sans serif faces. One solution has been to add serifs (often oversize ones) to the base of the figure, creating a character that is out of character with the rest of the font. Franklin Gothic is one such example. In a string of numbers the problem is unnoticeable since most figures do not have serifs in a serif font. But when set next to letters a seriffed 1 in a sans serif typeface jumps out like a sore thumb. One solution is to cut off the serif in Adobe Illustrator or FontLab and then kern it. Another is to use a more compatible 1 from another font—or an entirely different set of figures entirely.

top line: News Gothic bottom line: News Gothic kerned with baseline shift (hyphen).

Bulmer (Monotype, 1954; based on ATF Bulmer [1928] which was derived from the types of William Martin, 1792)
Finally, the lowly punctuation mark may be the flaw in a typeface. The exclamation point and the question mark are the most obvious of these since they are the largest. Monotype Bulmer has a chubby exclamation point, resembling an exploding cigar, that is at odds with the elegance of the rest of the typeface. It belongs in a Betty Boop cartoon. On the other hand, Weiss Antiqua has an exclamation point that is so short that it looks to be in danger of vanishing entirely. It is not a bang but a whimper.

Bulmer MT; Weiss Medium (tracked); and Century Schoolbook (tracked) for comparison.

Schneidler Medieval (F.H.E. Schneidler, Bauer, 1936)
To make the question mark in Weiss Antiqua equally short E.R. Weiss cut off the bottom portion where the character either shifts from a curve to a vertical line or simply curls up. Even stranger is the question mark for Schneidler Medieval. It appears to be upside down. For those who like the typeface, the question mark is its flaw. But for those who dislike Schneidler Medieval—and there are reasons, ranging from its overtly cupped serifs to its capital O—the weird question mark is no big deal.

top line: Weiss Medium and Schneidler Medieval (tracked) bottom line: FF Milo Serif for comparison; Schneidler Medieval (tracked and kerned).

After all, some flaws are only blemishes to those who deeply love a typeface. But other flaws are more serious as they hinder legibility, inhibit reading, or fail to perform their expected function. Whether the flaws in a given typeface are fatal or can be worked around depends not only on the typeface but on the typographer. Caveat littera!

35 thoughts on “Flawed Typefaces

  1. Pingback: Best of the Blogs: May 19, 2011 | CreativePro.com

  2. Pingback: Flawed Typeface « dogsmeat

  3. Paul Shaw

    Thanks to all of you who have supplied your own personal “flawed” typefaces or challenged me on mine. As Frederick said, these typefaces are not really flawed as in useless, they just have quirks that one has to adjust to. However, as much as I love blackletter (and can read it with relative ease) I do feel that antiqua is easier to read overall. The letter that gives me trouble is the k with its leg and arm turned into a bow so that it appears to be a bastardized t. I know what it is, but my eye stumbles over it each time. 

  4. Chris Donnelly

    Great article.I am actually more bothered by the “r” in the DTL Dorian image than the S or the s. (Actually, it’s the way the r feels so top heavy beside the o and the e.
    To me, it feels like a different weight altogether.
    It’s amazing how, when you really start to look at the nuances of type design,  beauty really is in the eye of the designer.  Thank goodness for that! It would be a shame to only have a handful of typefaces to play with.

  5. Peter Fraterdeus

    Paul, awesome as always. Wonderful to see such insight and careful observation of these atomic particles of the typographer’s art.

    As a printer, of course, my primary concern is whether type will ease into the press or fight with it.
    Many of the concerns you cite end up creating problems on press as well, particularly in small settings.

    I do think that the best type will always be that which has considered the nature of the medium it’s been designed for, whether screen, laser, planographic or relief plate.

    Best wishes from the heartland!!

    “zen of letterpress” cards

  6. Sigmatic

    You have to really love type to make it all the way through this article. There is truly a wealth of information here! Thanks for sharing (I made it about 33% and realized it wasn’t for me, but I really appreciate the value of your research).

  7. Matt

    When I say “alternate Y”, I mean that they use some sort of Y that isn’t the Palatino Y.  I don’t know if they made up their own or they borrowed a “Roman” one from another typeface.  But thanks for your reply.

  8. Michel

    Very interesting read! 🙂
    I think that you may wish to mention Century next time, too. I like Century but I think its dash (-) is flawed – it’s too high in the line… ( see sample ). I now wonder, why this is so? Is there a historical reason for this disabalanced position of the dash?…

  9. Paul Shaw

    Maybe these issues are more linguistic than cultural. Certainly, Mathieu brings up a good point that ampersands that more clearly show the e + t combination would be easier for French speakers to understand since et is “and” for them just as it was for the Romans. Similarly, Frederick’s ease with blackletter is surely helped by his fluency in German. (Though I cannot see blackletter as more readable than roman, despite my great love for it. It makes beautiful patterns and that is often antithetical to readability as there is less letter variation to aid the eye. I admire Zentenar but would not want to read a lengthy text in it. On the other hand, Jessenschrift or Wallau would be fine.)

  10. Mathieu Pigeon

    A few of these issues might be cultural. Maybe because French is my first language, neither the ligatured “Th” nor the Univers ampersand bother me. The “Th” is much rarer in French — unless one really love théâtre or théologie. As for the ampersand, also less common in French, it actually feels more legible, since it is recognizable as “Et” (“And”).

  11. Frederick D. S. Marshall

    I am a fan of Jean François Porchez’s Sabon Next. As he responded above, it is based on one of the three original Sabons – the Stempel, the least compromised and truest to Tchichold’s intentions in many people’s opinion. In my small palette of typefaces, it is my Beatrice Warde crystal goblet, my workhorse, the one I prefer as my typographic baseline for its ability to carry any kind of text effortlessly.
    As for Galliard, I’m with you; I’m glad Matthew Carter included the alternate, but I prefer the pelican-jawed g. Although the overwhelming majority of faces are too eclectic to carry most text, at the same time too many readers are unwilling to work at all for a text, to give themselves the chance to grow used to a face’s character. Galliard’s idiosyncrasies are organic and integral, and they make sense; the longer one reads Galliard italic, the more properly transparent it grows, though it has a nice bite at first.
    Finally, as for blackletters, although they take more getting used to, I find the best of them grow surprisingly readable if I give myself the time to get used to them. I’ve come around Hrant Papazian’s position that eventually, once we’re immersed in them, they are even more readable than the Latins are, that we only assume the Latins are more readable because of our familiarity with them. Even at tiny sizes, Gabriel Martínez Meave’s Darka is a delight to read; I like to set Hegel in Darka and read him in German.

  12. Paul Shaw

    Matt, there is no alternate Y in Palatino. The normal Y—one of the key letters in identifying Palatino or its sibling Aldus—has historical precedent behind it but often looks odd to modern eyes. It is a Greek Y or palm Y. The Romans did not have Y as a letter in their alphabet but the Greeks did. In the Renaissance Humanist scholars and scribes (most notably Bartolomeo Sanvito) revived the Greek Y as a sign of their erudition. It is a Y with curving arms (hence the palm designation) rather than angled strokes with serifs at the end. It is a difficult letter to design (many of Sanvito’s Greek Ys are poorly balanced) and one that does not fit comfortably into a line of caps because it has no clearcut top line. Sanvito and his contemporaries often avoided that problem by making it taller than the other capitals (following the Roman practice of having some Is taller than others). Other fonts besides Palatino (and Palatino Nova) that have Greek Ys include Aldus (and Aldus Nova), Michelangelo, Waters Titling, Sursum, LTC Forum Title, Caesario, Imperium and Requiem. The latter offers a Latin Y for those who are too timid to use the Greek one.

  13. Matt

    I’ve noticed that Mayo Clinic uses Palatino as their serif’d typeface. Someone must not have liked the way the capital Y looked when used in the brand logo, and replaced it with some kind of homebrew Y.  This is fine, except that I don’t think most people have access to that alternate Y…which causes some interesting inconsistent signage.  Sometimes there is inconsistency even within a single sign.

  14. Pingback: Best of the Blogs: May 19, 2011 | Themelabz.com

  15. Gray

    Avant Garde Gothic’s (and Lubalin Graph’s) uppercase R is by far my biggest pet peeve in terms of flawed typefaces. Such a beautiful, simple, good-for-anything typeface in all other regards, and then there’s that R that sticks out like a sore thumb. It’s gotten to the point that I won’t use it if there’s an uppercase R involved in what I’m making, and I’ve even considered making my own alternate version of it with a proper, correct R.

  16. Pingback: Flawed Typefaces | Sinting Link

  17. Paul Shaw

    John, script or calligraphic typefaces are a can of worms. I have been working on a book about them for the past three years. I just spent 8 hours sorting out the glyphs in Bickham Pro. But you are right about often not having the right capital letter. I usually find that I want a “plain” one and the only choice is an overly swashy one. I just came across that with R in His Nibs by Nick Curtis.
    Ambrose, if you have no problem with the Burgundica capitals then more power to you. I can read blackletter typefaces with little problem, but that is because I have years of experience with blackletter calligraphy, but most American designers I know stumble over letters like A, D, S and lowercase k and x in such typefaces. That is why I saw a bookjacket in the 1980s for that Adolf Hikler. 
    The idea of using roman capitals with blackletter seems weird at first but it actually has a solid history with German calligraphers and typographers. Rudolf Koch made it part of his Wallau and Jessenschrift typefaces and Hermann Zapf has often done this in his calligraphy and book design. And Noordzij designed Tret, a roman face, to match Burgundica. Unfortunately, it is not commercially available.

  18. John Brandt

    Oddly, the “Flawed Typefaces” link within my copy of the email didn’t work. I had to search the site. A fluke of email clients or browsers? Greatly enjoyed the article, even with minor disagreements. I applaud your willingness to so criticize heralded, famous(ly flawed) typefaces. I’ve often wished each new version (whether competing foundry or technology-based) would include alternates for such problematic glyphs, which I’d think would satisfy historical purists and those seeing incompatible or “difficult” characters. Earmarks and personality are one thing—downright illegible or extremely odd are one reason Fontographer was such a godsend initially, as such issues could suddenly be most easily addressed, especially when a vital usage contained the “flaw”. One addendum I’d be curious if others view similarly—which rarely would be considered a flaw—but something I’ve often found analogously vexing occurs often in many beautiful calligraphic swash scripts. In so many cases, inevitably the one swash cap needed is the only one in the set drawn most ordinary—unembellished and without characteristic ornamental swashes and swirls. Fortunately, ever more are released, easing the selection issue to a degree without the need to draw what for me are the most difficult to draw glyphs.That said, I appreciate and overwhelmingly agree with your postscript against advocating uniformity. I think most type junkies would agree that there can never be too many typefaces. A sizeable number will never see body text, but virtually any quirk can eventually find its unique use.

  19. Mark Sheppe

    The email that sent this article to my inbox today included the following “more” link:
    It was a no good. The link used on the title of the article, “Flawed Typefaces” did work. You might want to resend the email blast or redirect the bad link to the correct address. You’ll get the hits and responses you might be expecting.
    Good article btw.

  20. Paul Shaw

    Mike, I agree that many of the flaws that I have cited are earmarks for these faces. But earmarks are not necessarily positive aspects of a typeface. They are only characters that are distinctive enough to stand out and be easily remembered. I use the j in Centaur in my History of Type class as a key letter along with the M and R of Syntax.
    I am viewing these characters as flawed because they interfere with good typography as articulated by Beatrice Warde and others. That is, text typography that is invisible, putting as few obstacles as possible between the reader and the author’s words. I still believe that the Crystal Goblet approach to text typography is a valid one for over 90% of the things we read today. Although it is often seen as old-fashioned it is perfectly in sync with the attitudes of Swiss Modernism and contemporary attitudes in information design. Typefaces that disrupt reading, even if a little, with letters that act as tics and hiccups are not optimal. Why should we have to learn to accept the quirks if those quirks can be redesigned with no diminishment of the quality of a typeface? 
    The more we have to read the more we have to choose what to read. And one criteria is to choose those things which are inviting to read, are pleasurable to read and do not hinder our ultimate goal: to learn something, to be transported to another world, to be able to carry out a task, etc. Typography is a tool not an end in itself when it comes to most text. The Winter 2011 issue of the Bulletin of the Metropolitan Museum of Art is set in FF Quadraat. It is a handsome face that I found enjoyable to read—except for the R which continually caught my attention. Knowing it was there did not make it easier for me to accept. Instead it was like having a grain of sand in my shoe. Yes, I could still read the magazine but it would have been more pleasant with a different R.
    The flaws in these typefaces may be endearing quirks if they are being used for display typography or logo designs. Then, characters that stand out are desirable.  
    I am not advocating homogeneity or uniformity in typefaces. Far from it. I think we have too much of that today. Several of the typefaces I cited have quirky characters which I do not consider flaws since they do not hinder me when reading text set in them. The peculiar f and s in Optimo Didot the Elder help it stand out from the other Didots. The b, d, p and q were unique features of Gills Sans that set it apart from rival sans serifs in the metal era. These features have since been copied by many digital fonts such as FF Dax. Weiss Antiqua has an upside down S that I always like, though Stanley Morison didn’t. One of my favorite typefaces is Pegasus by Berthold Wolpe which is full of idiosyncratic letters, yet it works beautifully for text.
    Finally, some of these flaws go beyond subjective notions of readability. The default 1 in Gill Sans is a serious problem for anyone using that typeface for signage or information design purposes in which numbers play a significant role. 

  21. Joe Treacy

    Great article, Paul. You’ve shown how very easy it is for even some of typography’s greatest talents to make perplexing design choices that can have negative reverberations for decades. In designing typefaces, walking the line between providing unique design and ideal letterform readability versus a character becoming vanilla, overworked or inappropriate, is always an interesting balancing act.

  22. Bolen High

    Whether one agrees or disagrees with your take on the respective type-face flaws, it was a terrific article. While I deal with type every day, I don’t have your eye, and your carefully thought-out examination has helped shapren my eye and given me an appreciation for each of the mentioned font’s unique characteristics. I’ll never see an “I”, a “1”or an “i” the same again.

  23. Beth Koch

    The pairs of typefaces that you’ve compared are, to me, vastly different from each other. Searching for unique characters helps me as a designer to create a unique communication or logotype.

  24. Mike Yanega

    As someone who spends a lot of time trying to help others identify typefaces, mostly as a regular member of the Typophile Type ID Board, I had never thought of the features as flaws. To me they were earmarks.

    There is another aspect of this article that rather troubles me and that is the notion of tending towards a standardized idea of typographic beauty, if you want to use that word. Just as we get to know and love the faces of our family and friends, with all their deviations from the classic notions of beauty, so do we come to know and appreciate the typefaces that tell us stories, or the news of the day. We quickly learn to accept the quirks and read on with undiminished pleasure.

  25. Pingback: Flawed Typefaces | Middle Digit Media™ | Print, Design, Photography & Web Development | Miami, FL.

  26. Paul Shaw

    Barry, the Centaur j does indeed fit. I am sure that Rogers made it that way to avoid kerning problems. It is one of the letters I use to identify Centaur, but I still think it is not ideal. Of course, beauty is in the eye of the beholder. Bembo was not based on Griffo’s roman in the Hypnerotomachia Poliphili (1499) but on his roman in the 1495 De Aetna where indeed it does have a long leg—and causes spacing problems.
    Thanks for pointing out that P22, through their Lanston Type Company subsidiary, offers Goudy Oldstyle as it should be. The font is LTC Goudy Oldstyle Family Pro.
    Jean François, thanks for the corrections on my comments about Sabon Next. I checked the q/Q as a possible small cap but it came up in my glyph set as “Latin small letter q” and followed a swash lowercase n. I apologize for the mistake. I thought it looked a trifle large in the test word.

  27. Barry Schwartz

    LTC Goudy has the long descender, which does make Goudy better as a book face, no doubt, but would I see it all over the place in other roles if it had kept a longer descender?
    IMO the Centaur “j” looks that way because it looks great and fits in. Of course it may look awful with any given Centaur revival.
    Bembo is an Aldine revival, not simply some Garalde. Presumably the leg of the Bembo R is long because that’s how Griffo cut his R, not because it is elegant for titling purposes. See for instance the Posner copy of Hypnerotomachia: http://posner.library.cmu.edu/Posner/books/book.cgi?call=853_C71HY

  28. Jean François Porchez

    Paul, thanks for this wonderful article, interesting as always with your work.
    Just two details to be corrected about Sabon Next.
    1. Sabon Next Regular and Italic are strictly designed from Stempel version, nothing else. Careful printing of the glyph set, careful scan of all letters, then careful outline design. Printed, Sabon Next Display versus Stempel version looks the (almost) same at 12 pt.
    2. Sabon Next alternate q doesn’t exist. You used instead the Q small caps in your example. (Short tail Q, like short f exist on original version exactly like that.)

  29. katja

    “It belongs in a Betty Boop cartoon.” I also very much appreciate the suggestion to use an appropriate serif in the place of a Burgundica cap. Milo serif works so well!
    Great article.