GFS Heraklit

Daniel Benjamin Miller

Hello everyone, I'd like to share a new font which has been in the works for a while now. GFS Heraklit, to be released soon by the Greek Font Society, is a digital revival of Hermann Zapf's "Heraklit", which was designed in the 1950s as a Greek font to accompany the Latin Palatino. Although there have been some other Greek Palatinos released since then, none of them are digital versions of the Heraklit design, so this will be the first such digital font to be available. As far as I am aware, Heraklit has previously been released in two formats: manually-set metal type and phototype. We (George Matthiopoulos, Antonis Tsolomitis and myself) have developed this typeface to match contemporary PostScript Palatino (which is used for the Latin part of the font, in the form of an extended version of URW Palladio), and the subtle differences it has from the version of Palatino the original metal Heraklit was designed to accompany. We hope you enjoy.

Being a font from the Greek Font Society, GFS Heraklit, of course, supports polytonic Greek. It also has support for old-style figures and small capitals.

The sample attached does not have the final kerning and letterspacing, but the letterforms are pretty much done.

A release of the regular version is forthcoming (soon!), and two more releases, currently in progress, will expand upon that with a bold/bold "italic"/"italic" version (whether or not the word "italic" can apply to Greek I don't really want to debate right here, but these will be expansions upon, rather than revivals of, Zapf's design), as well as a set of mathematics fonts. But I don't have an exact timeframe for the completion of those parts of the project.

In the meantime, feel free to leave any comments on what we've got to show right now. I'm also interested in hearing your thoughts on any "italic" or mathematical design pointers for this style of typeface, and I will update this thread with some samples of those expansions when they're ready to share. Also, if you have any comments on the letterspacing/necessary kerning (besides the capital letters needing to be kerned with the lowercase, which is already in progress), please let me know what you think (see the attached PDF).

Nick Shinn

Small too bold.

Daniel Benjamin Miller

Nick Shinn said:

Small too bold.

The small caps are supposed to match the same design in (Adobe) Palatino, since this is supposed to be a Heraklit to match the digital-era Palatino design. (We don't use any of Adobe's outlines, though, of course.)

And I think the same "boldness" is exhibited here (compare the Bb here with the Bb in Heraklit, for instance), and that's the intended effect.
There is a different digital version of Palatino with more petite caps-sized smcp sets, but I did not want to go with petite caps. Maybe in the future that could be added as a different feature, but it would, of course, be significant time to draw another alphabet.

Nick Shinn

So, that being the situation (with the Latin small caps harmonizing with the Latin lower case), perhaps it’s the Greek lower case that is too light? 

Greek lower case tends to be a bit on the light side, when sharing certain parameters with Latin, due to its fewer serifs and vertical stems. 

Daniel Benjamin Miller

So, do you think the Greek lowercase is too light in comparison with the regular Latin lowercase? The Greek certainly has a lighter look than the Latin to some extent. But I think that's maybe due to the calligraphic nature. To me the shift doesn't look jarring printed out but maybe that's just me. But a possible worry is that they are not the same size, the Greek and Latin, exactly. I'll make an adjusted version…

Daniel Benjamin Miller

The beginning of my initial adjusted version. I think this looks more matching, yes?

Nick Shinn

I think one has to work with set text, observing Latin and Greek, Roman and Italic, U&lc and Small Caps in the same gaze, in order to compare and balance the overall weight of each, and to mix harmony and contrast evenly amongst the six styles and two scripts.

In particular, if your Greek lower case is too light now, that means your Greek Italic will be almost aetherial, if it takes its cure from Palatino Italic, which is so much lighter and narrower than the Roman. So don’t compromise the Greek Italic by developing it last, work it up in concert with its roman.

Daniel Benjamin Miller

I've redone the Greek letters; now, I think they mesh much better with the Latin.

Daniel Benjamin Miller

And a small sample which mixes Greek and Latin alphabets (maybe it should be μετὰ τῆς τῃ Palatino in line 4 but I digress).

Compare and contrast that text with these samples:

From Urworte der Philosophie 2nd ed. (1957):

And from the Phaedrus sample in the 1968 Manuale Typographicum:

(Note that we take our design for the capitals now from the 1968 version when there are differences, and that furthermore we adjust for, e.g., the new serifs on the Ε which were introduced in later Latin Palatino as well.)

Nick Shinn

Very hard to pick holes, nice job!

Minor points: I would try a little more demonstrative left-to-right downstrokes in several letters:

alpha—the right side looks a little thin, and it could be more pointy at the top, less round.

rho—the tail could be stronger

beta—the centre stroke could be stronger

kappa—stronger tail

tau—less hook, more emphasis on the downstroke—closer to iota

Similarly, the sigma-tau combination: in the original the sigma comes down on the tau,  rather than having a horizontal centre-line flow.

Of course, several of these features concern press gain (and pre-press gain), but we don’t have that now.

**

What’s with the position of the low accent in eta? 

Daniel Benjamin Miller

Nick Shinn said:

alpha—the right side looks a little thin, and it could be more pointy at the top, less round.

[…]

Similarly, the sigma-tau combination: in the original the sigma comes down on the tau,  rather than having a horizontal centre-line flow.

[…]

What’s with the position of the low accent in eta? 

The alpha actually is not round at its top right tip, but straight lines. But this could be made more obvious.

The σ and τ were set up as they were as part of the intentional design of the font: we're not going for a 100% authentic revival, but something that fits the spirit of contemporary Palatino (or, URW Palladio, which is the basis of our Latin).

And assuming you are referring to the iota subscript, this was a choice made to make GFS Heraklit fall in line with most Greek fonts. But we're not 100% that we'll go this way, and might choose the other way too.

Nick Shinn

Here’s what I meant about the “demonstrative” downstroke. The original alpha (right) has a little more angular ductus, and a right side downstroke weight which matches the left side curve. 

Yes, iota subscript was what I was referring to, I just forgot the name, haven’t done a polytonic font since 2008!

John Savard

Daniel Benjamin Miller said:

The small caps are supposed to match the same design in (Adobe) Palatino, since this is supposed to be a Heraklit to match the digital-era Palatino design.

And here I would have expected the small caps, like everything else, to match the typeface as it was embodied in metal type in the 1950s! However, either way, I am glad to hear it will soon become available.

Incidentally, if a companion version of a typeface is available which is slanted - but which isn't simply mechanically slanted - is there even any other word than "italic" to use to describe it, even if the character set is non-Latin? (And if such a word does exist, would anyone understand it if I were to use it?)

I mean, I guess I could say "with a slant" to say that the vertical axis of the letters are inclined without implying that the letters themselves were subjected to a process of being slanted...

Of course, if the purists that object to using the word italic in this case are Greek, then presumably they have coined a word to use in its place if one does not exist... which would mean that at least "the Greeks have a word for it".

Daniel Benjamin Miller

John Savard said:

Daniel Benjamin Miller said:

The small caps are supposed to match the same design in (Adobe) Palatino, since this is supposed to be a Heraklit to match the digital-era Palatino design.

And here I would have expected the small caps, like everything else, to match the typeface as it was embodied in metal type in the 1950s! However, either way, I am glad to hear it will soon become available.

Incidentally, if a companion version of a typeface is available which is slanted - but which isn't simply mechanically slanted - is there even any other word than "italic" to use to describe it, even if the character set is non-Latin? (And if such a word does exist, would anyone understand it if I were to use it?)

I mean, I guess I could say "with a slant" to say that the vertical axis of the letters are inclined without implying that the letters themselves were subjected to a process of being slanted...

Of course, if the purists that object to using the word italic in this case are Greek, then presumably they have coined a word to use in its place if one does not exist... which would mean that at least "the Greeks have a word for it".

My understanding is that in Modern Greek, it is called πλαγιά --- πλαγιός is an adjective for slanted and here's it's in reference to a γραμματοσειρά (typeface) πλαγιά  (slanted) … so they just call it "slanted" or "sloped". But my Greek knowledge is more classical than modern so maybe there are other words for this as well.

Of course the real reason it can't be called "italic" is because "italic" refers simply to a particular style, and this doesn't really apply to Greek. But there's no reason to think that "slanted" should mean "mechanically slanted".

I know from my talk with Antonis Tsolomitis that he isn't a "purist" in the sense of Greek font purists you expect. In his opinion the "purism" mostly has to do with non-Greek views of how Greek fonts should be, and he has no opposition to Greek fonts being like "Latin" fonts and so on and having an italic style — see his Kerkis fonts for an example, where he has designed the Greek to match the Latin italics of Bookman.

But what I can tell you about the italic GFS Heraklit — or ἡ πλαγιὰ GFS Heraklit, if you will — is that it won't just be a simple slant. But I do think that it won't be entirely dissimilar from a slanted version. I've used one as a placeholder of course but the results are entirely unsatisfactory so it's not worth sharing. Basically we'll do what we feel looks good. But we've not really designed the italic face yet. But since the upright is a revival, essentially, we felt that it was best to get that done first. And we think probably (though it's not 100% official) the release will be in three stages: first, we will release the regular font; then, we will finish with bold, italic and bold italic; then we will release a mathematical font (which is already in progress, but math fonts take a long time, so please be patient!).

As for the small caps, maybe I am biased, but I don't mind how they look at all. You will notice that the small caps in newer versions of Palatino (these are similar to those in PalatinoLTStd) are smaller and I rather prefer these to any petite caps as well.

But I think the feeling of the font is that it has to do with the whole of how we see the Digital Palatino — which, canonically, would be PostScript Palatino — and a dissatisfaction with the Greek designs which have been put forward for it. Even though Zapf approved, it's not my favorite anyway. So the kind of purpose of the GFS Heraklit project has been to bring the Heraklit design into accord with the Palatino design.

For instance, look at this sample. This is the same blurb I wrote about GFS Heraklit in Greek (I fixed a few typos/errors I found actually, which I fixed, but I digress). But now it is set in Palatino. I just find this very ugly and a poor fit with the Latin. But maybe that's just my opinion. Tell me whether you think this or GFS Heraklit is a better fit!

But at the same time a "classical Palatino" is not lost on me and I have some mind for doing that in a future project. So you heard it here first! Actually, it's been done before, but only in commercial contexts, whereas I would want to do a free and open-source font. According to Luc Devroye Softmaker Palazzo Original is based on this.

But, back in the very early stages of this font, George Matthiopoulos designed Greek capitals to match 1950s Palatino. Look at this:

Nick Shinn

Yes, the Palatino Greek shown is spotty and spindly in weight, and its angle of stress is inconsistent.

The lack of press gain is an issue.
It is probably OK used as a contrast style to Latin, in the traditional “Greek is Italic” manner, but mixed up in running text, it doesn’t work well. GFS Heraklit, being closer to the original, is much better.

Now, let’s see your “italic”!

Daniel Benjamin Miller

The (regular) face has been completed and its publication by the GFS is imminent (with samples, etc. as well).

John Savard

Daniel Benjamin Miller said:

My understanding is that in Modern Greek, it is called πλαγιά --- πλαγιός is an adjective for slanted and here's it's in reference to a γραμματοσειρά (typeface) πλαγιά  (slanted) … so they just call it "slanted" or "sloped". But my Greek knowledge is more classical than modern so maybe there are other words for this as well.

In that case, if your understanding is correct, we have a "correct" word to use in English now. We can call such typefaces "plagal", as this is already a word in English, even if it is used only in the field of music and not typography.

Normally, if a melody is written in the key of C major, it starts with a C and it ends with a C. And at the end, the chord progression will be a "cadence", going from G to C. At the end of a phrase within the song, a lesser cadence, going from F to C, may occur, and this is called a plagal cadence. The harmonic relation of G to C is that G is a perfect fifth above C, ideally, in just intonation, a 3:2 frequency ratio. The relation of the G to the next higher C is a perfect fourth, a 4:3 frequency ratio (since the ratio of one C to the next is an octave, a 2:1 frequency ratio.)

F is the note that is to C as C is to G. So F is a perfect fourth above C, and a perfect fifth below the next C.

I suppose that is the reason why it was thought appropriate to call that sort of cadence a "slanted" cadence.

John Savard

Daniel Benjamin Miller said:

But there's no reason to think that "slanted" should mean "mechanically slanted".

If there's a history of the term 'slanted' being used for mechanically slanted typefaces for the Latin-alphabet, then confusion might arise.

If what is being discussed is something that is close to what italic is for the Latin alphabet, one possibility that avoids the use of the term 'italic', but is also easy to understand while avoiding confusion would be to call the type weight 'inclined-cursive'. And for other cases, I suppose 'inclined-variant style' or something like that might do.

Nathan Zimet

This is a good revival i want to recommend to people who need a Greek typeface, but some of the thin strokes on the Greek letters might be too thin. On my lower res screen they get erased and make it hard to read.
I think calling the slanted style whatever word you think the customer would recognize/look for makes the most sense. 

Daniel Benjamin Miller

John Savard said:

Daniel Benjamin Miller said:

But there's no reason to think that "slanted" should mean "mechanically slanted".

If there's a history of the term 'slanted' being used for mechanically slanted typefaces for the Latin-alphabet, then confusion might arise.

If what is being discussed is something that is close to what italic is for the Latin alphabet, one possibility that avoids the use of the term 'italic', but is also easy to understand while avoiding confusion would be to call the type weight 'inclined-cursive'. And for other cases, I suppose 'inclined-variant style' or something like that might do.

I think the common terminology would be to call it "italic" in English, and πλαγιά in Greek. Antonis Tsolomitis seems to agree.

Nathan Zimet said:

This is a good revival i want to recommend to people who need a Greek typeface, but some of the thin strokes on the Greek letters might be too thin. On my lower res screen they get erased and make it hard to read.
I think calling the slanted style whatever word you think the customer would recognize/look for makes the most sense. 

What letters do you have issues with?

John Savard

Daniel Benjamin Miller said:

I think the common terminology would be to call it "italic" in English, and πλαγιά in Greek. Antonis Tsolomitis seems to agree.

I am perfectly in agreement with that. However, if there are those who feel that 'italic' would be inappropriate, given the difficulty of finding a good substitute, I thought it was still worthwhile for me to attempt to help finding something which had a chance of still being understood.