I actually favor the calt approach over an “fi” ligature approach, not just because it works better in Turkish but it works better in general.
I think for fonts I agree utterly and completely for the reasons you gave and because the solutions are more robust in everyday terms eg with letter spacing etc. In the case of handmade letters I am more undecided about it.
I can imagine a font where the ogonek shape might vary if two were used in succession for the same reason you might have a varient CALT f form to come after a first f. I would do the same thing ( exccept upside down) I think, make the second one a bit taller/deeper. But I don't know if such words exist. I'll have to look into it.
Thank you, Adam for the FF Meta example!
(Anyway, I feel like I have spammed the thread with my previous post. If I have, I am sorry).
I feel like I have spammed the thread
No, you haven't. Your post was well within the Typophile guidelines. :-)
I can imagine a font where the ogonek shape might vary if two were used in succession[...]
There is no possibility of such succession (in polish). Ogoneks describe "nasality" (ą sounds like the end of french mon and ę like the middle of cinq. It is even not possible to spell eg "ąą" easily. More: none polish word begins with Aogonek nor Eogonek.
"none polish word begins with Aogonek nor Eogonek."
Thanks, Marcin! It is very helpful to know this kind of information when kerning. Not just in Polish, but all languages. I have a particular need with the dcaron, lcaron, tcaron abutting other diacritic glyphs. How often do such collisions occur in real use and what are the common combinations?
I was making a list of tasks yesterday and I started to get distracted by drawing an aogonek next to other glyphs to see what I thought the impacts/interactions, if any might be. The a seemed like to most likely because The letters I thought about were g,j,y. The q's descender is too far to worry about. agonek and uogonek with g & y seem like the most likely to interact. But looking at Adams designs I see that you can and probably should work around that with a better/more robust design instead. Maybe there might be subtle color/eveness issues that might be addressed with a CALT though. untimately I have to learn more about common patterns of use as Chris suggests.
The pairs ąg, ąp, ęg, ęp are pretty common combinations in Polish. I cannot really say so much about other languages that use the ogonek (well, I do have some data but...)
It is true that Ą or Ę never starts a word in Polish so kerning pairs like Ąw are nonsense. But typically, you'd use Ą as a dependent kerning glyph in a class driven by the A glyph, so the overhead of Aw -> Ąw is minimal. Except, of course, you expand the class pairs and generate pure singleton pair-based kerning, say, for a Type 1 CE-encoded font.
Thanks Adam! That helps.
Adam, what you posted make me wonder, could you see a ĄJ or Ąy combination? or a Qį or QĮ?
Aogonek and eogonek appear with low frequency in polish text. (My statistics locate aogonek at 1% and eogonek at 1,3% when a (the winner!) has 8,5% and e - 7,6%).
Aogonek preceds c,g,p,t,z,zdotaccent
letters preceding aogonek (most often): c,i,j,k,n,S
Eogonek preceds b,d,g,k,p,s,sacute,t
letters preceding eogonek (most often): i,j,Lslash,n
As you can see, ą and ę kern with some similar signs; you can also kern them interchangeably (with all neighbours from both groups, because ą & ę are sometimes exchanged, eg in singular/plural). You could also consider: Tą, Tę. Aogonek and eogonek can appear at the end of a word.
Eben, Ąj and Ąy combinations don't exist in polish, but also probably nowhere at all (I checked both in Google: 0 results; just some spam-like, artificially generated sites).
Marcin, thanks for checking!
Two ogoneks in a row can occur in Native American languages: there ogonek signifies nasalisation, and doubling the vowel signifies length. According to Wikipedia long nasalised vowels exist in languages like Apache or Navajo and the grapheme for them is (as expected) doubled vowels with ogonek.
EDIT: This is a comment for Eben's question of her ogoneks descenders depth, from 1st page of the thread. I posted the answer because I forgot there were more pages...
Your ogoneks seem to descend more than "g" and "y" and that - in my opinion - is a bit too much. You can correct it by thining the shapes (which are fine anyway), as you have planned.
Eben is a he.
He may or may not have heard a Who.
My fault, Eben! But let's leave the post untouched: it sounds so politically correct ;)