List of glyphs to omit

Michel Boyer

I just downloaded the OpenOffice German language extension and unzipped it to check for fj and then, out of curiosity, I looked for fi. The hyphenation dictionary contains 33 entries with an f-i hyphenation including

chef1ingenieur
grund1stoff1in1dustrie
kauf1impuls
kunst1stoff1in1dustrie
prüf1in1genieur
tropf1in1fusion
zupf1in1strument

Should I understand that, for example, the fi ligature in kaufimpuls is incorrect? If so, does InDesign do anything to prevent it? (I have no working version to check). And if not, how can you make a font that can be used for German without disabling the ligature for German?

Nick Shinn

Halfback (a position in American football).

Ray Larabie

Offbeat is a common English ffb word. Other fb words: snuffbox, halfbeak, wolfberry, puffball, surfboard, hoofbeats, goofballs, surfbirds and surfboat. And Kafkaesque for fk.

Max Phillips

Hofbrau.

Thanks for the Maltese lesson, Kent. That's great to know. I'm still not drawing f_f_j, though. Or f_f_b. You've got to draw the line somewhere, and dipping snuff is bad for you.

In other news, I've quit supporting Esperanto. They're component glyphs that take no time to draw, but they'll never be used and ĥ is just too damn ugly.

Jens Kutilek

Michel Boyer said:

chef1ingenieur
grund1stoff1in1dustrie
kauf1impuls
kunst1stoff1in1dustrie
prüf1in1genieur
tropf1in1fusion
zupf1in1strument

Should I understand that, for example, the fi ligature in kaufimpuls is incorrect? If so, does InDesign do anything to prevent it? (I have no working version to check). And if not, how can you make a font that can be used for German without disabling the ligature for German?

That’s right, all of these ligatures should be suppressed. InDesign does nothing to prevent it, but there is a script available on Friedrich Forssman’s website to fix that (scroll down all the way to find it).

Michel Boyer

If I understand correctly, the script uses grep to select the offending ligatures, turns off liga on the selection and then selects metrics kerning. Neat.

Also, from the following comment, I understand that InDesign could already handle some exceptions

// die folgenden werden von Indesign automatisch richtig interpretiert
// "[ff]-[fi]",	// Kraftstofffilter
// "[ff]-fa",	// *schifffahrt*, Zellstofffabrik, ALT: Hofffahrt, ...
// "[ff]-fä",	// Trefffähigkeit
// "[ff]-fo",	// Kunststofffolie
// "[ff]-fr",	// Hafffrau, holzschlifffrei, schadstofffrei, ...
// "[ff]-[fl]",	// Auspuffflamme, Sauerstoffflasche, Schlifffläche, ...
// "[ff]-fü",	// Stofffülle

The ... are mine so that everything fits in the rectangle.

Michel Boyer

The ligatures in the rectangle above are treated correctly only if the font has no f_f_f ligature (which would be exceptional indeed); by default InDesign does nothing special but to apply the ligatures in the font when activated.

Johannes Neumeier

Maybe I'm stating the obvious here, but the way I draw those precomposed f-ligatures is via components anyway, so it is very little work to give the liga feature a class substitution for any f followed by a lowercase stem. I often have regular f, f component for f_f, f component for f_i/f_j and an f component for f_stem. With some fonts f_f and f_stem can be the same, or f_i/f_j does not need a f component at all. Also the f_i/f_j component f can sometimes be useful for handpicked f_accent combos. Then you don't need to think how obscure combinations you want to support or not, but the feature code handles it all.

Michel Boyer

I don't see how you can decide at the glyph level for an fi ligature in  pressechefin and no fi ligature in chefingenieur

Christian Thalmann

I thought the prohibition against such ligatures in German was a largely historical and not very official rule. Ligatures are simply supposed to be solutions to a typographic problem, not linguistic helpers. We're used to reading nonbreakingcompoundwords in German, after all; what harm is the ligature going to do? Do people even consciously notice ligatures when reading? If so, did the ligature fail at its job?

On a related note, I generally put /f_f_* ligatures in my typefaces because I enjoy them, but in my latest project (a sans), they tend to look a bit too monolithic. Is it preferable to allow the /f_f ligature to trigger and just use non-ligature coping mechanisms to avoid collisions?  I guess /f_f_i should still be included, but in the other cases, the /f_f part will pretty likely constitute the end of a morpheme and thus welcome the separation from the following part, right?

Michel Boyer

I was always fascinated by an italic f used in Springer's Ergebnisse der Mathematik. Here is from Behnke & Thullen, Theorie der Funktionen mehrerer komplexer Veränlichen, Zweite erweiterte Auflage, 1970.

    

It seems to me the editors were for some reason particularly afraid of ligatures.

Max Phillips

I generally put /f_f_* ligatures in my typefaces because I enjoy them, but in my latest project (a sans), they tend to look a bit too monolithic.

There are always non-joining ligatures.