Greek alphabet?

[RGB-es]

While I understand the main goal of this font is for medievalists, it would be great for math teachers to have the Greek alphabet (and maybe more math symbols). Is there any plan for adding this?

Thanks!

[psb1558]

This comes up every now and then. I love the Greek alphabet, and I would love to have it all in Junicode, but when I sit down to actually draw the letters, I find myself bewildered, because I have little feeling for that script, and the idea of producing a Greek alphabet that harmonizes with Junicode's Latin is daunting and overwhelming.

So when this has come up in the past, I've suggested that users select one of the wonderful free fonts of the Greek Font Society. For harmonizing with Junicode I especially love GFS Bodoni and GFS Baskerville. These do a terrific job of honoring the designs of the past while also serving present needs.

And of course, Junicode is Open Source, so anyone crazy enough to take this on is free to do so. Fork the project; design the Greek alphabet; send me a PR! Your name will be immortalized on GitHub among Junicode's contributors.

[psb1558]

On the other hand, I did try my hand at Greek some years ago—a close imitation of of an 18th-c face.

Maybe I should try (again) to adapt it for use in Junicode.

[RGB-es]

Looks great!

[psb1558]

Lowercase roughed in. Still need to work on spacing and getting a consistent color. Capitals not yet complete. As this is a historically oriented font, I suppose all the polytonic stuff had better be included.

The italic lowercase won't be much different: just a few more degrees of slant (and it is, so far, less complete than the roman):

[psb1558]

With revised kappa (the script or old-style kappa now at U+03F0).

[RGB-es]

Awesome!

[marrus-sh]

If you are doing Greek (I completely understand not wanting to 😅), would it be possible to also get stigma (Ϛϛ) / digamma (Ϝϝ) / koppa (Ϙϙ/Ϟϟ) / and sampi (Ͳͳ/Ϡϡ) for their use in Greek numerals? (Also U+0375, and probably U+0374, although the latter is canonically equivalent to U+02B9.)

[psb1558]

My plan is to make the Greek sections complete. So yes, all those things.

[psb1558]

Roughed in now, with all the axes in place. Spacing is obviously a mess, but I'll see to that before a new version is out.

Regular:

Condensed Bold:

[KSundstrom]

Would this then require anew the use of mgrk in order to access harmonized variants of the most commonly used lowercase Greek letters in e.g. linguistics (IPA) and maths, or would rather the current ss17 Linguistic alternates come into play for both linguistics and maths, or do the relevant characters perhaps have different codepoints for use in linguistics and maths (as opposed to Greek proper)?

[psb1558]

I've been so preoccupied with the letters themselves that I haven't thought much about presentation. I think I saw somwhere (but I didn't bookmark it and now can't find it) that mgrk was deprecated, or was going to be. In any case, it seems to be out of favor. Instead, for math, you've got (1) the Greek symbols spots in the Greek range (for a limited set of characters) and, more comprehensively, (2) the Unicode Mathematical Alphanumeric Symbols block, which can contain a variant for every letter of the Greek alphabet.

So mathematics is well taken care of by Unicode, and I suppose that's why mgrk is going away. (I know next to nothing about mathematics, and I've never tried particularly to make Junicode a math-friendly font. I do intend to correct the basic math characters for this font the way I did--with a lot of helpful input from users--for Elstob some time ago.)

For linguistics (which doesn't seem to have dedicated code points for Greek variants), different approaches are possible. One is to use ss17 (which I formerly used for a single phonetic symbol, now on a cvNN feature) for all the Greek letters for which linguists expect a different shape. This is the approach the Brill font takes (but it uses ss20). (I'm now using ss17 for something else, but I've pretty much decided to retire that, so ss17 could be used for phonetic alternates again.)

Another approach would be to make the romanized forms needed to harmonize with the other IPA characters the default and use locl to shift to the Wilsonian Greek forms when the script and language are Greek. This approach would require no action on a user's part, but would handle the business silently—as long as the text has the proper language tags.

I'd like to hear from people what approach would be best.

[psb1558]

Junicode now has its own Greek face, newly designed to harmonize with Junicode roman. It is in the roman face only: the italic still has the old Wilson Greek.

There's still a fair amount to do: fixing some letter-shapes, fixing metrics, new kerning. But to my eye, anyway, it's now at least in the "could be worse" category.

You can try it out in the current build. Comments invited.

[psb1558]

(2 hours later) snuck in a new roman with a few fixes.