Message-ID:  <9702251338.AA27210@epsilon.ams.org>
Reply-To: Mailing list for the LaTeX3 project
              <LATEX-L@RELAY.URZ.UNI-HEIDELBERG.DE>
In-Reply-To:  "Your message of Mon, 24 Feb 97 22:26:18 +0100." 
              <3312077A.3B72@sssk.se>
Date:         Tue, 25 Feb 1997 08:38:13 -0500
From: Michael Downes <mjd@MATH.AMS.ORG>
Sender: Mailing list for the LaTeX3 project
              <LATEX-L@RELAY.URZ.UNI-HEIDELBERG.DE>
To: Multiple recipients of list LATEX-L
              <LATEX-L@RELAY.URZ.UNI-HEIDELBERG.DE>
Subject:      Re: International documents
Status: R

> The thing is that - in Swedish - compuond words in which the first part
> ends in a double consonant when a word by itself and the second part
> starts with the same letter get this letter in triplicate when
> hyphenated.  Example:  "tillaga" (make (food)) is hyphenated "till-laga"

Knuth discusses this for German "Bettuch" -> "Bett-tuch" in the TeXbook,
exercise 14.8.

This kind of behavior in LaTeX is limited by the capabilities of the TeX
typesetting engine. TeX cannot do it, unless you write "tilllaga" and
use fonts with a suitable ligature spec to collapse the three l's into
two. Compare the "ffi" ligature in "sufficient" where TeX collapses
three letters into a single character; the separate letters will be
resurrected if it is necessary to break at that point.

Or you could write the double-triple consonants with a LaTeX command
such as

  \DeclareTextCommand{\ll}{T1}{ll\discretionary{-}{l}{}}
  ...
  ti\ll aga

The use of \discretionary will have the side effect of suppressing all
other hyphenation in the word. Naming these might be a bit problematic
since \ss, for example, is already used for German eszet.

In general, if some instances of "ll" should be normal and some should
be latent triple "lll", TeX will not know which kind is which unless you
distinguish them by some difference in the input. This might not be
necessary if TeX hyphenation patterns had a way to specify such
behavior---but they don't. (If they did, paragraph breaking would
undoubtedly take longer than it does now.) Using Omega instead of TeX
might be a solution...

Michael Downes, mjd@ams.org