My view as a multi-script designer (if not an expert in either Greek or Cyrillic) is that as a rule one should try to make so-called "cognate" characters somehow different, irrespective of whether there's any historic reason for a given glyph to diverge. This is because: information (in this case concerning what language/script one is reading) comes from contrast; and it gives a given script a stronger identity.

One of my favorite examples is the "Y" in Adobe Sava.

BTW, Armenian has a similar glyph, the Piwr Փ.

hhp

Why?
Because you don't have to make them the same.
If your design process includes assessing a glyph shape by setting it in text in the Metrics window, then the different scripts might lead you to optimize Ef and Phi differently. I ended up making Ef wider than Phi, in one of my typefaces, because it looked better that way in words.

Consider capital Lamda and El, for a drastic example of differentiation.
In Greek, Lambda is always triangular, but in Cyrillic it may be either triangular (in "old style" and certain geometric styles) or rectangular (the "modern" style), depending on the genre of type.

The Ef with extenders beyond cap boundaries is somewhat mannered, as evidenced by its adoption at the HermesSoft foundry as the pointedly different Bulgarian Cyrillic style.

As with Latin (think of Q's tail), there is room to maneuver if you are designing an original typeface. There are all sorts of variants of Ef, two-bowl, one-bowl, with and without prominent "extenders".

> I ended up making Ef wider than Phi

Width is actually a very relevant factor here. Some scripts* need looser setting - and naturally this affects the "inside whites", and potentially the black bodies as well. Admittedly there is relevance to uniformity, but really only in display setting (and not even always) and pretty much none in text (except in very broad stylistic terms). Blindly making so-called "cognate" characters the same is just bad design in my book.

* Like Armenian. And really any two scripts (in fact any
two languages!) necessarily have different spacing needs.

hhp

Again I forgot the big donkey:
Most of all -in the more important lowercase- you probably don't want the x-height (and to a lesser extent any of the vertical proportions) to be indentical, at least in a text font system.

See: http://typophile.com/node/81755

hhp

Thank you gentlemen. But do you also make A B C E H I M N P O S T X Z in Cyrillic or Greek differently from Latin or each other?
In many cases, the outlines (and even the sidebearings) of these characters are completely identical.

I guess that I don't need to do it because Ф is rather difficult letter to harmonise with other letters because of the density, and there are different ways to do it; other than that, common letters like A and B can be perfectly identical.

But even with those which do not include extended Phi Φ and Ef Ф, they are still designed differently.

My guess is this: Cyrillic and Greek versions of Latin typefaces are often done by different designers at the same time, so they may end up with different solutions for similar shapes.

But do you also make A B C E H I M N P O S T X Z in Cyrillic or Greek differently from Latin or each other?

I have seen that too, but I don’t remember exactly in which typefaces. If somebody is adding a different script, they ideally shouldn’t hesitate to change existing letters when adapting them for the new script if required.

> In many cases, the outlines (and even the sidebearings)
> of these characters are completely identical.

Yes. Let's fix that.

hhp

> In many cases, the outlines (and even the sidebearings) of these characters are completely identical.

> Yes. Let's fix that.

Or, let's not 'fix' that.

I'm not necessarily comfortable with saying it is a noble objective to make like glyphs physically differ from each other in the name of respecting cultural diversity. Now, I'm not suggesting that you shouldn't be free to make like glyphs differ in both outline and metrics but no-one should imply that it is 'wrong' to use exactly the same data, if to do so works as well or, dare I say, better than using a different shape simply because the option is there.

> Why?
> Because you don't have to make them the same.

Nor do you have to make them different, notwithstanding your sensible reasons for doing so. There may be equally good or better reasons for keeping them the same (for example, because keeping them the same works!)

> If you are aware of this issue, please tell me why they are treated that way (historical or practical reasons behind it), and the problem of using identical outline.

There may not be a 'problem'. Surely a good way of harmonising across languages is to identify similar characters at the outset and making the rest of the characters harmonise with these like characters? Cue howls of derision and mocking because I don't know about a particular ductus! Well, apologies to the purists.

First off: I for one can't be accused of "ductus fetishism". :-)

Concerning having a reason to break apart cognates, as I've said elsewhere, it depends where a given design is on the Display-Text axis (noting that it can never be at either extreme). Formal harmony of shapes is admittedly very relevant towards the Display end; but in my mind there can be no doubt that the information arising from differences handily overpowers such considerations towards the Text end.

hhp

I agree with you both, but I actually don't feel the need to change simple characters like A or T, even in text faces. It seems to go against the idea of typeface: uniformity (of course everyone has a different sense of uniformity and diversity, so the sweet spot is vast). At least I will keep it that way until I am fully aware of the differences of each script (it's only the first time for me to design a multi-script typeface). Anyway, my Φs in my new text face are individually adjusted to fit Greek and Cyrillic!

>hrant
Out of curiosity, I'm wondering how Armenian Piwr should be made differently. If you make a typeface that consists of Latin/Greek/Cyrillic/Armenian how would you do to those three Φs? If the answer is too easy, let's say the Armenian should be 100% vertical as the others.

> I'm wondering how Armenian Piwr should be made differently.

The Piwr is indeed difficult to diverge. I frankly don't know how to make it structurally different without also making it too peculiar (at least in a contemporary context; I could use a strange historical form as a pretext, but I'm not into that sort of thing). However there's always the factor of width, which is almost guaranteed to have different needs between scripts, due to:

1) The ideal overall letterspacing, which affects not just the sidebearings but the inside whites as well.
2) The typical/average width that the letters want to be. For example Armenian capitals (in the contemporary context*) are narrower than Latin overall because there are many arches and virtually no diagonals.
3) In the case of capital letters, the necessity to stand out from its lowercase (assuming the cap height is nicely lower than the ascenders). For example when the lowercase is particularly tall (many Armenian letters span the entire height) its capital needs to be extra wide. For example ֆ vs Ֆ. The փ is also pretty hefty, so I make the Փ wider than I otherwise would.

* Armenian caps also have the archaic option of diverging greatly in width.
See: http://typophile.com/node/66620

hhp

Cyrillic Ф and Greek Φ in Trebuchet…
…and in ITC Charter

Go figure.

One thing is clear, though: those glyphs are—most often—different in either shape, or size, or both. It is customary in Cyrillic type design to let the stem of the Ф project above the cap-line. And yes, in some Cyrillic faces it also drops below the baseline—sometimes quite considerably, like in Telingater Display.