Path: utzoo!utgpu!jarvis.csri.toronto.edu!cs.utexas.edu!wuarchive!brutus.cs.uiuc.edu!jarthur!dhosek From: dhosek@jarthur.Claremont.EDU (dhosek) Newsgroups: comp.text Subject: Re: National alphabets Message-ID: <4527@jarthur.Claremont.EDU> Date: 21 Feb 90 09:52:51 GMT References: <1261@shelby.Stanford.EDU> <1064@kunivv1.sci.kun.nl> Organization: Pitzer College, Claremont, CA 91711 Lines: 34 In article <1064@kunivv1.sci.kun.nl> scribitur: >In Dutch the 'ij' is really a ligature: the lip of the 'j' >should reach under the 'i'. Very close kerning! Let's consider >this a national character. In doing research for a type design project, I noticed the ij ligature appearing in all sorts of contexts, but never with anything more than "i-j ligature" to identify it. Now I at least know the language responsible for this. Also, in every typeface that I saw this combination occur, it appeared that the i and j of the ligature were identical to the ordinary i and j. The ligature status appears to be an artifact of a time when the close kerning described wasn't possible (think about TeX's boxes and imagine they can't overlap and there's your typesetting problem with ij). Now, does ij-lig occur every time i+j appears or only sometimes? (in Latin, for example, \oe is only used if oe represents a dipthong. Thus p\oe{}na, but poeta.) If it occurs every time, then an implicit kern would be more efficient than a ligature (they both need a different TFM from CM, but the former needs no additional glyphs). If it occurs only sometimes, \def\ij{i\kern-j} would work but would prohibit hyphenation (the accent/hyphenation problem isn't because of accents, per se, but because TeX does not hyphenate words with explicit kerns), so in this case, having an ij national character would make sense. -dh -- "Odi et amo, quare id faciam, fortasse requiris? nescio, sed fieri sentio et excrucior" -Catullus D.A. Hosek. UUCP: uunet!jarthur!dhosek Internet: dhosek@hmcvax.claremont.edu