home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!mcsun!uknet!edcastle!edcogsci!iad
- From: iad@cogsci.ed.ac.uk (Ivan A Derzhanski)
- Newsgroups: comp.text.tex
- Subject: Query: kerns, accents, ligatures
- Message-ID: <11396@leonard.ed.ac.uk>
- Date: 9 Nov 92 14:55:51 GMT
- Organization: Centre for Cognitive Science, Edinburgh, UK
- Lines: 48
-
-
- (1) TeX doesn't kern accented letters even if it kerns their
- unaccented counterparts. The file "roman.mf" says
-
- ligtable "V": [...] "A" kern kkk# [...];
-
- and accordingly if {VA} is set in Computer Modern, the "A" goes 2u
- under the "V". But in {\'VA} or {V\'A} no kerning is done.
-
- Is this a bug or is it a feature? That is, is kerning in the case of
- one of the letters having an accent deemed typographically undesirable,
- or is it just something that we can't get with the available tools?
-
- (2) I've frequently encountered the opinion that if one has to
- accommodate an alphabet with more than 26 characters (separate glyphs
- or accented forms of one another), the way to go is to use ligatures,
- not active characters. Y Haralambous suggests in _TuGboat_ 10:3 that
- (_inter alia_) {<A} be made into a ligature which yields "A" with an
- acute accent. So {V<A} will be set as {V\'A} while also allowing for
- hyphenation later on in the word.
-
- But, as I understand it, there will still be no kerning. The ligtable
- entry for "V" says that a following "A" must go under it, but it can't
- say anything about a following "<", because it is still unknown what
- comes after the "<". When that becomes known, in the entry for "<",
- where it says `"A" =: A_acute', it is too late to move the ligature.
-
- Or am I missing something?
-
- (Apparently it is considered a bad strategy to put the modifiers after
- the letter, because in that case {fi<} won't yield the desired "f"
- followed by "i" with an acute, if the "fi" ligature is still used.
- But it would also be a bad idea if our ligatures are not accented
- characters, but shapes which look completely different from the
- components. For example, if a Cyrillic _sh_ is entered as {sh} (or
- {s1}, or {s<}, whatever, via ligature building), it will be kerned
- into a preceding letter as if it were an _s_, which is wrong, because
- the two look entirely different from the left side -- the former must
- be kerned as a Roman "m" and the latter as a Roman "c".)
-
- (Yes, folks, I'm writing MF code for alphabets with more than 26
- letters, and I want them to look good. ...)
-
- --
- `Haud yer wheesht! Come oot o the man an gie him peace.' (The Glasgow Gospel)
- Ivan A Derzhanski (iad@cogsci.ed.ac.uk; iad@chaos.cs.brandeis.edu)
- * Centre for Cognitive Science, 2 Buccleuch Place, Edinburgh EH8 9LW, UK
- * Cowan House, Pollock Halls, 18 Holyrood Park Road, Edinburgh EH16 5BD, UK
-