The kgreenlandic (kra) has a lowercase form only. Good question as to what to use in a small caps setting (or in an all caps setting, for that matter). We'll need an expert to weigh in, but other than avoiding such a setting, my inclination would be to design a usual kra glyph that matched the height of the small caps (typically slightly higher than x-height) and another to match cap height. But that might be seen as incorrect by native users. Just looking at Minion Pro, I don't see anything other than the normal x-height kra. The kra is now replaced by "q".

Thank you, Gary. As far as I know both letters are not being used any more.
http://www.eki.ee/letter/chardata.cgi?ucode=0149

But if you add them to your font, people should at least be able to use it correctly. And when they switch to small caps and these letters would stay lower case, that would not be quite elegant ...

I don't have any answer, but the problem applies to some other characters as well:

• 01F0 - LATIN SMALL LETTER J WITH CARON
• 1E96 - LATIN SMALL LETTER H WITH LINE BELOW
• 1E97 - LATIN SMALL LETTER T WITH DIAERESIS
• 1E98 - LATIN SMALL LETTER W WITH RING ABOVE
• 1E99 - LATIN SMALL LETTER Y WITH RING ABOVE

All these characters don't have a uppercase mapping in Unicode.
What should happen when the [smcp] feature is applied? Are they 'small capsed' and if - do they keep or lose their accents?
What should happen if the user switches to 'All-Caps'? Do i need a uppercase variant for the [case] feature as well?

 According to Michael Everson’s The Alphabets of Europe, http://www.evertype.com/alphabets/greenlandic.pdf, the capital associated with [ĸ] is [K‘], with a “high-6” single open quote. (Take that, SmartyPants!) Presumably a small-caps K followed by a (lowered?) open quote would be appropriate.
 As concerns the letters eigi mentions, code points like U+01F0 [ǰ] are considered compositions (eg., of j + ◌̌); the small-caps version should just combine the small-caps j with U+303C Combining Caron. Abusing the IPA small-caps, it should look something like [ᴊ̌].

Thank you, JC, I forgot to look at Everson's site.
Meanwhile I had a look at Arno Pro in InDesign CS3 using the buttons for small caps and all caps:

The problem shown here is that InDesign is not applying Unicode special casing rules, so doesn't know what to do with these diacritic letters when case conversion is applied.

Thank you, John, so what would you recommend? I think adding at least the small caps would not be the worst idea ...

If the Kra has a distinctive form as we see in the example above ( no?); making a special small cap shouldn't be too hard ... But the design issue & the technical one are seperate. John, you are saying that Indesign should be doing something. What do you think of putting code in to point to the correct Special glyph. Isn't that what happens All caps & Small Caps anyway?

I have been increasingly convinced that a seperate Small cap font rather than a single font containg Roman Caps & Small Caps is a good idea anyway since Indesign & other apps don't have a very elegant way of handling the issue yet - and may or may not evolve one. Do you agree?

 SIL and Everson actually disagree here; the one claims upper-case Kra is K’ (K + U+02BC modifier letter apostrophe), the other, Kʻ (K + U+02BB modifier letter turned comma—which is supposed to look like K‘ i.e. K + left single quote; Georgia seems to display it as K+prime). At any rate, the modifiers are the preferred characters to append, not the quote characters.

—Joel

This is so obscure you could probably set your own standard.
There are a few archaic Greek "single-case" letters also: historically, only one form was ever used, but Unicode, being a two-case system (yo Eben, don't get me started) calls forth the other.
So the question is, who is going to design the glyph?
Do type designers leave it to the orthographers, who will likely (e.g. Everson, SIL) just add some kind of tick mark? (Pardon me, "diacritic":-)
I say, if we have to make the damn things, we should at least be allowed a little fun and invention!

Good catch, Joel!

Samuel Kleinschmidt, who introduced the pre-reform Greenlandic orthography (and the one responsible for the kra, if I understand correctly) in his 1851 Greenlandic grammar, doesn't seem to have cared about capitalization. I've skimmed the contents at Google Books, and nowhere does he seem to use capitals for any Greenlandic words. His German is far more sparse on capitals than I am used, for that matter.

I tried to find examples of Greenlandic books in the pre-1973 orthography at Google Books, and found this example which has the form K’ânâĸ (Qaanaaq in today's orthography) with what looks like a modifier letter apostrophe. So score one for SIL. This is just one example, though, and it's a rather small sample size. I would like to see more, but it's difficult to find a fully browsable book written in pre-1973-orthography Greenlandic at Google Books for copyright reasons.

I did find this snippet of part of a map with "K’ânâĸ" written by hand, but it's too small to see what's really going on in there with the diacritical marks.

Nick, the example with the capital kra that I just posted is from a 1909 book. So this issue is likely something that conforms to its own set of conventions, and Everson and SIL are not arbitrarily inventing anything. They are merely describing what seems to have been standard practice.

Nick: This is so obscure you could probably set your own standard.
This is a whole language and a literary tradition you are talking about. Do you think type designers should disregard the accepted practices just to have fun in designing a new glyph, which would only serve to cause confusion in the representation of historic spellings? At least the German eszett is still in use so the possibility is there for a gradual acceptance of a capital form.

The archaic Greek letters pre-dating the bicameral divergence of the Greek alphabet are a fun challenge for type designers, because there is a need for them, as you said. That's where we can exercise creativity and work toward building up some kind of consensus. The Kra doesn't need our help.

I would suggest eliding the diacritic with the letter.
That would avoid the mark bumping into the diacritic on a following letter.
Also, changing the nature of the top right serif from the standard upper case form of K (borrowing from, say, C) would provide consistency with the lower case treatment. I mean, the lower case character isn't just a "k" with a diacritic, is it?

But really, what possible use is there in retconning a deprecated glyph?

Especially if the "retconning" isn't done to fill a gap but to replace something that was completely legitimate in the past. In that case, I guess it's not really retroactive continuity at all but simple revisionism.

Yes, you're right. Retconning is done to resolve conflicting accounts of the same event into a single historical reality.

So this is more like anti-retconning, suggesting an alternative universe/history that has continuity with the present day--referring to historical continuity, not continuity between the cases as mandated by Unicode.

Really it's just historical fiction, although I do like the Marxist overtones of "revisionism".

In type design, a comparison might be the Renaissance style of Petrine Cyrillics, or Storm's 18th century sans serif, Sebastien.

When you try to offer justification for going one route (SIL) or another ( Nick's) you end up flailing to a certain extent if you insist on grabbing something solid/absolute. You could take either position legitimately as long as you acknowledge your reasons for doing what you do. My preference would be to look to the historical precedent - and then step out from there if at all. But doing that is only my prerogative - not the only right way.

It is important to consider that a language with not a ton of readers & not a a ton of books was probably given less thought & consideration typographically speaking than other languages. It seems likely to me that the glyph design solutions they evolved might be a bit too slapped together and could stand being reconsidered especially if it was done in a manner that was culturally historically sympathetic. Acknowledgment that a given solution is completely legitimate doesn't preclude new ideas. How do you know that the Kra couldn't be helped? I don't! The world can't be split into simply retroactive continuity vs. simple revisionism. Insisting on this split takes your eyes of glyph designs and therefore real consideration of the issue.

What is the population of readers of this glyph set anyway?

SIL and I may disagree, but I'm sure I have K‘ rather than K’ because that's what was in a source available to me.

Michael Everson
evertype.com

John Hinz's 1944 Grammar and vocabulary of the Eskimo language as spoken by the Kuskokwim and Southwest Coast Eskimos of Alaska (Bethlehem, PA: Moravian Church Society for Propagating the Gospel) uses the Greenlandic Kra. No examples of capitalization, alas.

Michael Everson
evertype.com

Interesting thread.

So… what should happen with napostrophe if we are using small caps or all caps settings?

It seems likely to me that the glyph design solutions they evolved might be a bit too slapped together and could stand being reconsidered especially if it was done in a manner that was culturally historically sympathetic.

Kra was phased out of the official Kalaallisut orthography in the early 1970s, so the only time one is going to find it used to today is likely when one is attempting to faithfully reproduce older works. It would seem to me that a reconsideration of its design would be at odds with that goal.

André

I think it is a mistake to place the apostrophe mark over the left stem of the n like this, Cristobal. The old Afrikaans character U+0149 is not a letter with a combining mark it is, as the Unicode name indicates LATIN SMALL LETTER N PRECEDED BY APOSTROPHE. It has a compatibility decomposition to U+02BC (spacing apostrophe modifier) followed by U+006E (n), and the Unicode special casing rules indicate that the uppercase form is U+02BC followed by U+004E (N). In other words, the typical rendering for this diacritic character should be identical to ’n and ’N.

Thanks John, I’ll follow your advise.