hyphenation patterns in language.dat: Why not enable all?
[email protected] (Frank Küster)
| Newsgroups | gmane.comp.tex.tetex.beta |
|---|---|
| Message-ID | <[email protected]> |
Hi all, it seems some users don't find it very easy to enable new language patterns (like forgetting to actually build the format etc.). In a discussion in de.comp.text.tex, Walter Schmidt asked why we wouldn't just enable all, or nearly all hyphenation patterns. My first answer was "Hm, bad things might happen, let's not do it". However, in fact it's only the size of the hyphenation trie which is increased (and each language uses one or a couple of multiletter control sequences). Therefore we (Walter and I) suggest to enable most patterns by default. I have made some tries: english, german, ngerman, french and nohyph: 14 hyphenation exceptions Hyphenation trie of length 23210 has 711 ops out of 35111 30 patterns (all included in March, except only one for norwegian and ukranian, respectively): Hyphenation trie of length 112338 has 3208 ops out of 35111 103 for language 30 [...] That makes an average of 107 per language, and we are still below 10%. The benefits of this, in my opinion, would be that users of all common languages would not need to bother about this at all, the hyphenation would simply be there. The only disadvantage I see is the larger size of the format files - in the test cases it was 1.6MByte vs. 750 kByte for latex.fmt, but tetex-2.0's latex.fmt is 1.3MByte with the four patterns only!. With today's disk sizes, I would not say this matters much. What do you think? -- Frank Küster, Biozentrum der Univ. Basel Abt. Biophysikalische Chemie