To go back to the starting point of this international documents discussion: The `Duden' book (see below for a description) mentioned before by in relation to math typography has an appendix `Richtlinien fuer den fremdsprachigen Satz' (guidelines for the typesetting of foreign languages), giving rules in 10 categories for 20 languages, namely (keeping the numbering) Categories: (1. name of the language) 2. abbreviations 3. capitalization (usage of uppercase) 4. quotation marks 5. numbers 6. date format 7. measures and weights 8. price labels (?, price statements, price formats? `Preisangaben') 9. postal code 10. punctuation Languages (as in alphabetical order in German): Danish, English, Finnish, French, Italian, Croatian, Latin, Dutch, Norwegian, Polish, Portuguese, Romanian, Swedish, Spanish, Czech, Turkish, Hungarian, Greek, Russian, Serbian. May be that list could provide a good working base to discuss the implentation of related features in LaTeX Some comments on the categories: 2. could be partly handled by TeX, or TeX could at least help the user here: in some languages, an abbreviation is always followed by `.', in others only sometimes (in English depending on whether in ends with the last letter of the abbreviated word). In German, the parts of `multi-part abbreviations' (things like u.A.w.g.) should be seperated by a thin space (\thinspace does this, but this could be offered to the user via a macro) 3. seems to be burdened on the user 4. could be handled via macros (logical markup of quoation marks instead of using ligatures) 5. may be partly, may be by a macro with to arguments stating the input format and the desired output format (e.g. as language names) 6. could well be handled, but Tex would need to know of the desired format (German? or Austrian? or Swiss?), and that's a matter of style or taste. 7.,8.,9. I can't see a way for TeX of knowing about these. May be by supplying language-sensitive control sequences which would complain about a wrong format, but there seem too much possible cases. The `Duden' referred to here is `Satz- und Korrekturanweisungen' Duden-Taschenbuch Band 6, Bibliographisches Institut, Mannheim: 1986 (5th ed.) I think it is out of print unfortunately, but I hope there will be a new revised edition soon. It is mainly based on assuming an author-printer relationship which is quite oldfashioned now, but it is still very useful. I hope that a new edition will fully acknowledge the computer, but may be DTP is just the reason why there isn't an up-to-date edition. Anyway, the book contains rules and symbols for markup of a manuscript or typoscript to be understood by the printer, also for corrections in the galley proofs and so on. More interesting are parts as the list above, a list of symbols given with reference numbers (quite useful for the communication with the printer in the old days, still useful as a source of many symbols not yet covered by TeX and as a reference for their main usage), rules for typesetting mathematics, physics and chemistry etc etc. I think this particular Duden book is less descriptive as the Duden related to German language, but of course it doensn't have the sense of typographical beauty and the instructiveness of e.g. Tschichold's writings, rather in collects rules and covers many areas of typography. And as for another matter which occured in the discussion a few weeks ago: In the Duden book are lists of transliterations and transcriptions for various languages, something which could be very interesting for TeX: at least transliteration could be well covered by TeX (ArabTeX has such a feature), I doubt this for transcription, here TeX could possibly only give a basic transcript to be treated by the user. To explain the terms here: transliteration gives a bijective map, a one-to-one relationship between glyphs, so this map is reversible. Transliteration is language-dependent: Russian transliterated to German isn't the same as transliterated to English, as the glyphs are chosen to give an idea of the pronunciation, and there could be different standards alongside for the same pair of languages (e.g. there are two or three for the transliteration of Chinese to German). But this could all very well be done by TeX, if the appropriate mappings are provided. Transcription is another thing: Here's no such bijection, sometimes one glyph is transcribed by a letter group, sometimes a lettergroup by a single letter (e.g. Russian `ks' as in Aleksandr is rendered `x' in German transcription), so this is not reversible neither easy for TeX, as sometimes the map is relating on context. Johannes Kuester -- Johannes Kuester [log in to unmask] Mathematisches Institut der Technischen Universitaet Muenchen