MFB Oldstyle (and ahistorical Greek)
Daniel Benjamin Miller
Posts: 126
Two years ago, I released MFB Oldstyle. I didn't post about it, since it is just an open-source (PD via CC-Zero) revival of Century Oldstyle (MFB = "Morris Fuller Benton"), and, while useful (I hope!), it was nothing groundbreaking in terms of design. The font is available on GitHub and CTAN.
This year, I released some updates. Besides finally finishing the bold italic I'd been working on (and making a few tweaks/fixes more generally), I also added Greek support to the font. The bold italic was not my own original design; while not part of the original release of Century Oldstyle, it was put out in later revisions and I in fact worked from the ATF specimens of it.
My Greek, however, is totally new. It's also, I would say, ahistorical. Not only was Century Oldstyle never released with Greek characters back in the day (I don't know about other modern digital versions, but that's not really my concern), it wouldn't have really fit in with the paradigm of the day in a number of ways. The typical practice when a text called for both Latin and Greek alphabets to be included was to just mix two typefaces; a book set in Century Oldstyle could set Greek words or passages in, say, Porson. The idea of getting the Greek to match the Latin within the same family wasn't really that common then, as far as I know.
Additionally, while there are plenty of slanted Greek fonts (I suppose "italic" could be said to be a misnomer if applied to the Greek script), I don't believe that the upright–slanted distinction would have been present in a single font family (or used in the same Greek text) until fairly recently, so that's another way in which this Greek typography is not in line with the common practice of the early 20th century.
That said, it's 2026, not 1909, and many font families have Greek across the four standard variants. Users might not necessarily expect it, but it's useful to have, and so, like many others who have done font revivals, I've decided to diverge from the source material (at least in this respect). I hope that my Greek designs are a good visual match (more or less) for my Latin ones; they are a bit different, in line with what I expect when I read Greek. They are not based on any existing design in particular, though you might see what my influences were!
The italic, I'll admit, is funky; standard slanted Greek would normally be more like Porson, and, indeed, I wouldn't really enjoy this design for extended passages in Greek, unlike the regular, but I wanted it to fit well with my Latin italic font.
I welcome anyone's thoughts on the design (though it's already been released!), and on the general points relating to adding Greek to an existing Latin typeface.
***
Anyway, here it is (as present in version 2.1 of the font). I hope you'll excuse the long-ish sample; here, you can see the alphabet, a Greek sample (from Aristotle) and two short English quotes with Greek mixed in (both from W. K. C. Guthrie).
Regular
Italic
Bold
Bold Italic
Tagged:
2
Comments
-
out of curiosity, what led you to choose CC0 over 0BSD, unlicense, or other public domain-equivalents? it's been a while since I've evaluated these kinds of licenses, but was recently thinking about 0BSD for a project just out of familiarity with other BSD licenses..0
-
jeremy tribby said:out of curiosity, what led you to choose CC0 over 0BSD, unlicense, or other public domain-equivalents? it's been a while since I've evaluated these kinds of licenses, but was recently thinking about 0BSD for a project just out of familiarity with other BSD licenses..Functionally, there is no difference between these. They all result in your work being usable without any restrictions whatsoever.To be clear, CC-Zero is not just a public-domain-equivalent license, but first and foremost a waiver, stating explicitly that the copyright has been given up, plus an extra fallback which gives a robust equivalent license just in case this is not legally effective (even though it should be most places if not everywhere, and certainly is in the United States, but the efficacy of copyright waivers is a subject for another time and place). This is why I say it's public domain via CC-Zero.If you want to put your work in the public domain ("or equivalent"), it comes down to your preference, but here's why I made this choice specifically:
- In the US (and some other places), typefaces are not copyrightable. Despite what many font people think (or want), I think it's pretty likely that digital outline fonts aren't actually copyrightable either (see also Matthew Butterick's comments on Laatz v. Zazzle). I say that I am disclaiming any copyright I may have, which means I'm explicitly avoiding making a copyright claim to begin with.
- Relatedly, the BSD-0 license says "this software." The Unlicense also uses the terminology of "software." While I don't think any reasonable court would interpret it restrictively, (a) it's arguable that outline fonts aren't properly "software" at all (see above), and (b) in any case, I want my dedication to be as explicitly broad as possible.
- The Unlicense in particular is written in a confusing and poorly thought out way.
- CC-Zero is a copyright waiver which is as comprehensive and explicit as they come, which is why it's the most generally recommended copyright waiver, as far as I can see.
- "Public domain" is a much clearer way of expressing what you can do than "BSD-0" or "unlicense." "Unlicense," in particular, is an extremely confusing name, because being Unlicense-licensed ("Unlicensed") is the opposite of being unlicensed.
In short, if you want to waive all copyright, why not just say so? That's what I wanted to do, so it's what I did.That being said, I'd prefer to keep the focus on the font design here, rather than make this thread about CC0 or licensing or whatever.2 -
that makes a lot sense to meIn short, if you want to waive all copyright, why not just say so?I have seen large projects/organizations that are explicit about which third party licenses are blanket approved to be used vs. require some kind of approval. so I guess "speed" and/or "risk-averse legal teams with other priorities" could be one somewhat flimsy answer to this question, but I imagine CC0 is well-covered these days. the shaders in vello are an interesting case where they both state what they want ("The intent is for this research to be used in as broad a context as possible") while also providing unlicense, MIT, and apache licenses, presumably so that such an approval process isn't necessary at organizations/in projects that are already fine with those licenses. but that seems a bit cumbersome to me, and I would imagine confusing for some end-users.
anyway, I don't mean to get off topic - thanks for sharing your reasoning, particularly around the specific lack of "software" language0 -
@Daniel Benjamin Miller Very interesting. With a few minor adjustments, it could even be turned into a variable font, allowing for the creation of intermediate weights.
1 -
Having the Greek lowercase follow the notional pen angle of the Latin always feels wrong to me now, which makes it difficult to look past at the details. One thing that stands out is the shallowness of the oxia and varia accents, which should be a lot steeper.There are some problems in relative proportions of letters: e.g. the uppercase Ξ looks massive, while its lowercase ξ — and also the ζ — feel pinched and too narrow. The spacing feels a bit uneven — also the Latin to a lesser degree — but that may be a result of the pattern of light and dark caused by the pen angle.0
-
Not a native user of Greek so take these comments with a grain of salt!
The head serif on roman /eta (and I suppose /kappa) strikes me as stiff and Latin-y. Is it a intentional link to the Latin?
All four styles feel lighter than the Latin counterparts. Were you trying for matching weight, or a little contrast?
The little thin turn south at the very bottom of italic /rho feels a bit brittle and out of scale to me.
Lowercase italic /psi looks too weighted towards the left side, and all the capital /Psi’s look quite dark.0 -
michele casanova said:@Daniel Benjamin Miller Very interesting. With a few minor adjustments, it could even be turned into a variable font, allowing for the creation of intermediate weights.
This is mostly true, although there are some differences between the regular and bold weights — particularly in the italic p. That said, I have never personally understood the appeal of variable-weight fonts. I have put thought into making this a variable font, but my goal would be to have an optical size axis, not a weight one, matching the original Century Oldstyle's availability in various optical sizes. To me, that would be much more useful.John Hudson said:Having the Greek lowercase follow the notional pen angle of the Latin always feels wrong to me now, which makes it difficult to look past at the details. One thing that stands out is the shallowness of the oxia and varia accents, which should be a lot steeper.I understand that it's not traditional, but it's also not uncommon now, and I personally like if it aligns with the Latin in this way. On the other hand, I find those Greek fonts which have been made to actually look like Latin hard to read. I guess it's a matter of taste.
Yes, everything is intentionally liked to the Latin. As for why they would perhaps feel lighter to your eye, I'd guess this is because of the contrast between the serifed Latin characters and the Greek lowercase design.Craig Eliason said:
The head serif on roman /eta (and I suppose /kappa) strikes me as stiff and Latin-y. Is it a intentional link to the Latin?
All four styles feel lighter than the Latin counterparts. Were you trying for matching weight, or a little contrast?
0 -
@John HudsonHaving the Greek lowercase follow the notional pen angle of the Latin always feels wrong to me now,That is the conservative point of view, but it’s never bothered me, and is perfectly acceptable in Greece today (see Parachute’s designs).
After all, the Roman/Italic system of emphasis originated in the Latin script, where both “true” italics and sloped romans are well established as legitimate, whatever the genre.
So if one is going to import it into Greek, both methods may be part of the “package.”
And the OP did say “ahistorical.”0 -
As I said, it feels wrong to me and now. I have spent too much time working with the traditional script and looking at texts — not just typography, but also manuscript, other print media, and handwriting —, so Greeks with Latin stroke modulation just look wrong, especially certain letters like λ where it is really obvious that the shape of the letter evolved with a different modulation pattern.Letter shapes are not neutral carriers of arbitrary stroke weight: they are the product of writing with particular tools in particular ways. Some shapes are more or less amenable than others to rearrangement of modulation patterns, and some really don’t work except as they evolved.I am quite interested in hybrid Greek designs, in which modulation is applied less rigidly one way or the other, and instead varies across the alphabet.0
-
I’m inclined to agree with you that the lambda is a bit too strong, with that great long heavy stroke (compared with, say, chi).
But perhaps if there were not the ball terminal at the top of the ascender, and if there were some kind of serif at the bottom of the thin stroke, the effect might be tempered. In other words, there are tweaks that can be made, just as in the Latin (e.g. lower case z with its “wrong” ductus).0 -
John Hudson said:Having the Greek lowercase follow the notional pen angle of the Latin always feels wrong to me now,I'm surprised to hear that. I thought that Porson, with the lower-case essentially italicized, while the upper-case was upright, is currently considered to be incorrect - which is why other Greek typefaces designed for setting Greek text, virtually without exception, avoid following its example.That being said, Porson, of course, does excel for one particular purpose - the setting of mathematics.
0 -
John Savard said:I'm surprised to hear that. I thought that Porson, with the lower-case essentially italicized, while the upper-case was upright, is currently considered to be incorrect - which is why other Greek typefaces designed for setting Greek text, virtually without exception, avoid following its example.John Hudson is talking about the pen angle, i.e., the way the strokes work, not the character angle.That being said, I don't know how Porson could be considered "incorrect." It's just not really used for modern Greek; it still appears frequently for classical texts.0
-
@Daniel Benjamin MillerI agree that the lowercase letter 'p' in the italic version would need to be modified to create a consistent variable font. I believe the appeal of a variable font lies in the ability to control the characteristics of a typeface using just two files (regular and italic) across one or more axes; for instance, using a slightly lighter weight for white text on a dark background compared to the same text in black on a light background.
This is mostly true, although there are some differences between the regular and bold weights — particularly in the italic p. That said, I have never personally understood the appeal of variable-weight fonts. I have put thought into making this a variable font, but my goal would be to have an optical size axis, not a weight one, matching the original Century Oldstyle's availability in various optical sizes. To me, that would be much more useful.
0 -
I was thinking of perhaps having a variable axis for ink spread (someone asked me for this), allowing for slightly spindlier or slightly inkier printing, but this would be separate from the boldness control, since it would be omnidirectional, whereas boldness is not.michele casanova said:@Daniel Benjamin MillerI agree that the lowercase letter 'p' in the italic version would need to be modified to create a consistent variable font. I believe the appeal of a variable font lies in the ability to control the characteristics of a typeface using just two files (regular and italic) across one or more axes; for instance, using a slightly lighter weight for white text on a dark background compared to the same text in black on a light background.
This is mostly true, although there are some differences between the regular and bold weights — particularly in the italic p. That said, I have never personally understood the appeal of variable-weight fonts. I have put thought into making this a variable font, but my goal would be to have an optical size axis, not a weight one, matching the original Century Oldstyle's availability in various optical sizes. To me, that would be much more useful.0 -
What you are describing there is usuallly referred to as a grade axis. The other key factor of grade is that it doesn’t affect spacing, so weight is added or trimmed without changing the advance widths of the glyphs.1
-
John Hudson said:What you are describing there is usuallly referred to as a grade axis. The other key factor of grade is that it doesn’t affect spacing, so weight is added or trimmed without changing the advance widths of the glyphs.Yes, exactly. The word had escaped me in the morning. I do feel like this is also the ideal approach for the use casde that Michele brought up.0
-
Daniel Benjamin Miller said:John Hudson is talking about the pen angle, i.e., the way the strokes work, not the character angle.Oops. Not reading carefully on my part.I was so shocked by what I thought I was reading, I stopped before looking carefully.Pen angle is something that characterizes a typeface stylistically, so for Latin or Cyrillic, it comes as a matter of course that the pen angle should be the same for the upper-case and the lower-case. One wouldn't want to mash up what would be essentially the lower-case from one typeface with the upper-case from a completely different typeface of a completely different style.But, of course, that doesn't apply to Greek, for obvious reasons. I don't know if having the same pen angle for both is "wrong", but definitely there's no reason why having a different one for each would be wrong.As for Porson: I'm not a native speaker of Greek, so I can't really be sure. But given that pretty much every other Greek typeface I've seen doesn't follow Porson, the fact that classical Greek texts printed in English-speaking countries may be set in it could just be down to habit, and the availability of fonts with the necessary apparatus for rendering polytonic Greek.I mean, if Porson didn't horrify the native speakers of Greek who encountered it, one would expect several other Greek typefaces to have the characteristic of upright upper-case and italic lower-case applied to other styles of letter. Instead, this seems to have been, as far as I can tell, never to have been emulated by anyone else.0
-
John Savard said:As for Porson: I'm not a native speaker of Greek, so I can't really be sure. But given that pretty much every other Greek typeface I've seen doesn't follow Porson, the fact that classical Greek texts printed in English-speaking countries may be set in it could just be down to habit, and the availability of fonts with the necessary apparatus for rendering polytonic Greek.I mean, if Porson didn't horrify the native speakers of Greek who encountered it, one would expect several other Greek typefaces to have the characteristic of upright upper-case and italic lower-case applied to other styles of letter. Instead, this seems to have been, as far as I can tell, never to have been emulated by anyone else.Well, I am not a native speaker of Greek, but I did study classical Greek, and encountered a number of modern Greeks along the way, none of whom ever said a harsh word about Porson. Then again, these were scholars, not typographers, and they never said anything to me about fonts at all!I would say that the Porson font definitely comes off as old-fashioned, but that doesn't make it wrong or horrifying; it's just something that you probably wouldn't see much in modern Greek texts today.As far as I know, all the people involved with the Greek Font Society (except for me) have been native speakers of (modern) Greek, and they put out a version of Porson, so clearly they don't object to it. Their blurb on Porson, in its Greek text, says that the Porson font was in use in Greece until the advent of digital technology. (That said, anecdotally, I'm pretty sure it was less commonly used in Greece than Didot-style fonts.) Yannis Haralambous, in "From Unicode to Typography, a Case Study: the Greek Script," includes an example of Porson being used in a Greek children's book in the 1960s, which he describes as "typical."I would suspect that the decline of Porson comes from the digital-era expectation of multiple styles (matching the four styles in a standard font family), which discouraged the major computer typeface firms from shipping a Porson-style design (none of them do for system fonts, as far as I'm aware). And this also leads to the contemporary Greek practice of borrowing "italic" typography from Latin, etc., which precludes the use of Porson in the everyday context.0
-
Daniel Benjamin Miller said:As far as I know, all the people involved with the Greek Font Society (except for me) have been native speakers of (modern) Greek, and they put out a version of Porson, so clearly they don't object to it.But none of the other typefaces featured on their site follows Porson. So this isn't what Greeks think is how a Greek typeface should work, or so it seemed to me.The version of Times New Roman that came with my copy of Windows doesn't seem to include Greek. The Greek lowercase is upright in Liberation Serif and in TEX Gyre Termes - the latter surprised me, since one major purpose of TEX Gyre Termes is typesetting math, or so I thought.I would suspect that the decline of Porson comes from the digital-era expectation of multiple styles (matching the four styles in a standard font family), which discouraged the major computer typeface firms from shipping a Porson-style design (none of them do for system fonts, as far as I'm aware). And this also leads to the contemporary Greek practice of borrowing "italic" typography from Latin, etc., which precludes the use of Porson in the everyday context.I absolutely agree that this is a possible cause, or at least a contributing factor. Since the GFS features a wide selection of Greek typefaces from various historical periods, though, the Greek rejection of the Porson dichotomy for their own typefaces seems to have been around before the digital age. Of course, though, people have been using italic for emphasis in hot metal, and so the Greeks may have wanted to do that too long before the digital age.0
-
This is taking us off the OP’s original topic. But since Daniel himself engaged on the Porson topic …Daniel Benjamin Miller said:I would suspect that the decline of Porson comes from the digital-era expectation of multiple styles (matching the four styles in a standard font family), which discouraged the major computer typeface firms from shipping a Porson-style design (none of them do for system fonts, as far as I'm aware). And this also leads to the contemporary Greek practice of borrowing "italic" typography from Latin, etc., which precludes the use of Porson in the everyday context.When Font Bureau was commissioned to provide Miller Banner fonts for a since-abandoned Microsoft product, with a full WGL4 charset (which includes Greek), I was asked to help work on the Greek component.Matthew had drawn a separate Porsonian font on a previous occasion, and it just made historical sense to use that as a starting point to reunite with the Austin-inspired Miller.
Even though the MS commission was only for Banner, we opted to sketch out Text first as a foundation, and in case of future development across the family. I started by adjusting the proportions, weight, and angle of Matthew’s Porsonion glyphs to marry up with the Miller Text Italic.
From there, we adapted an ahistorical (if you will) upright version — somewhat the opposite of the usual Latin practice of designing a secondary italic to pair with the roman.
And from that, we derived the more extreme Banner styles.
I don’t believe the Microsoft project ever really went anywhere. And I’m not sure what’s become of the fonts now. (Although, of course, I still have an archive of my sources.)Neither Matthew nor I are Greek experts; but we were pleased enough with what we did under the circumstances.3 -
The version of Times New Roman that came with my copy of Windows doesn't seem to include Greek.Microsoft’s TNR definitely includes Greek, and has done so for decades. It’s not very nice, but it is there.The stroke modulation in the lowercase is hybrid, and shows reasonably convincingly how that can be done.
0 -
John Hudson said:Microsoft’s TNR definitely includes Greek, and has done so for decades. It’s not very nice, but it is there.
You're absolutely right. I don't know how I managed to scroll past it (in Character Map), and yet see the Cyrillic.0 -
I welcome anyone's thoughts on the design (though it's already been released!), and on the general points relating to adding Greek to an existing Latin typeface.
I've seen better greek fonts, TBH. Even in preinstalled system fonts - I'd say Cambria has a good neutral, legible greek. Book Antiqua has a highly legible greek yet not too modern-looking.
So I am wondering why I'd not just use one of those? Not to put down your initiatives, whatever they are. But from the user perspective, if I were to read a lot of greek text (which on its own is rather hard on eyes), why I'd not simply pick a better one "for free".
In short, the one in yours don't feel like a mature font. Mostly I think it is due to odd stroke contrast, combined with this kind of Minion-ish stroke style, create rather unpleasant optical effect.
IMO the very idea of adjusting something to look similar to Latin counterpart does not have much practical and theoretical sense. Weight and relative size adjustment makes sense. Probably playing with proportions.
How one can adjust something to look similar to Latin if one don't know what that "something" should be in the first place?0 -
John Savard said:I mean, if Porson didn't horrify the native speakers of Greek who encountered it, one would expect several other Greek typefaces to have the characteristic of upright upper-case and italic lower-case applied to other styles of letter. Instead, this seems to have been, as far as I can tell, never to have been emulated by anyone else.
At the risk of veering wildly off-topic, I’ll point out that in modern Greece, there is a noticeable—though not universal—trend toward abandoning lowercase letters entirely. For instance, I have a perfect example right at hand: a design featuring several different typefaces, overlaid with a handwritten inscription—all in uppercase.
2
Categories
- All Categories
- 47 Introductions
- 4K Typeface Design
- 501 Type Design Critiques
- 593 Type Design Software
- 1.1K Type Design Technique & Theory
- 675 Type Business
- 908 Font Technology
- 29 Punchcutting
- 544 Typography
- 129 Type Education
- 333 Type History
- 82 Type Resources
- 114 Lettering and Calligraphy
- 33 Lettering Critiques
- 81 Lettering Technique & Theory
- 580 Announcements
- 101 Events
- 116 Job Postings
- 176 Type Releases
- 186 Miscellaneous News
- 270 About TypeDrawers
- 54 TypeDrawers Announcements
- 114 Suggestions and Bug Reports