hyph-utf8 update (2010-03-21)

git-svn-id: svn://tug.org/texlive/trunk@17577 c570f23f-e606-0410-a88d-b1316a301751

1 files changed, 11 insertions, 0 deletions

diff --git a/Master/texmf-dist/doc/generic/hyph-utf8/sa/hyphenmin.txt b/Master/texmf-dist/doc/generic/hyph-utf8/sa/hyphenmin.txt
new file mode 100644
index 00000000000..5fbd790b657
--- /dev/null
+++ b/Master/texmf-dist/doc/generic/hyph-utf8/sa/hyphenmin.txt
@@ -0,0 +1,11 @@
+This is the explanation of the values of hyphenmin used in Sanskrit (and possibly Indic languages) as given by Yves Codet in a private mail on March 19, 2010:
+
+For lefthyphenmin I think 1 is correct: in Sanskrit, if we allow breaks such as "tu-dati" ("tu" = 2 characters) we must allow "pa-tati" ("pa" = 1 character).
+
+For righthyphenmin I first suggested 5, as François wrote, which should be forgotten (and forgiven :). Then I thought 3 could be good; my idea was to prohibit a break before a consonant at the end of a word, such as "apata-t" (final "t" = 2 characters). But if hyphenation patterns have precedence over hyphenmins, since patterns contain rules prohibiting such breaks, that precaution is useless, and it has a drawback: it forbids breaks such as "pata-ti" ("ti" = 2 characters) or "pata-ta" ("ta" = 1 character). If patterns do have precedence, it now seems to me that the best choice is 1 for both hyphenmins (I'm sorry, François, about my hesitations :).
+
+But if my reasoning is sound, this would be for Sanskrit and Prakrit, in whatever script they are written. Maybe there would be reasons to choose other values for some other Indian languages, which I don't know.
+
+Kind regards,
+
+Yves