From: Ralf Stubner <[EMAIL PROTECTED]> > On Sun, Jan 07, 2007 at 03:28 +0100, Danai SAE-HAN wrote: > [ UTF-8 and TeX ] > > There are packages like utf8x that are under development, but IMHO > > that just takes too much time and we need a quick fix. UTF-8 has been > > around for many years now. > > Actually utff8x (or more appropriately ucs.sty) are no longer under > development. At least I haven't heared of anyone responding to the > original author's orphaning. IMO the real way forward does not come from > clever macro programming but from extensions to the underlying TeX > engine that support Unicode. In principle omega has been able to do that > for quite some time. Unfortunately almost nobody knows how to use it. > However, now there is XeTeX that has native Unicode (and OpenType) > support. And the pdfTeX successor (luaTeX) will have similar abilities.
Indeed, newer TeX engines like Omega/Lambda have Unicode support. But I miss a good introduction document. Apparently, you just continue using LaTeX and add a few "OCP" and "OTP" commands at the start of your document, but that seems rather easy to get used to once we have good examples. More important is the issue of fonts: you have to convert to Omega Virtual fonts, and I have no experience with that. Vincent Zoonekynd has an article about Japanese with Omega; I'll check it out later on http://zoonek.free.fr/LaTeX/Dictionnaire_japonais/main.html and http://zoonek.free.fr/LaTeX/FontInstallUnicode/documentation.html. The latter website talks about getting Bitstream Cyberbit to work with Omega. Although I haven't tried it (yet), I see that his method includes PK fonts. OMG! But perhaps this is because he uses the old method with ttf2tfm. CJK on LaTeX now uses another mechanism with Type1 fonts; I wonder if this can also be applied on Omega. If so, then I might be able to use my current Type1 font packages. This shows, however, that if we are going to use Omega/Lambda, it will be even more complicated than is now the case: not only do we have to create the fonts (like I do now for example for CJK on LaTeX), but we also have to make OVP files. His article about Japanese dictionaries also discusses XML input, so it might be useful for debiandoc-sgml. Robert Wright shows us on http://pws.prserv.net/Roger_Wright/ROGER.HTM how to use it with (classic) Chinese, but I'm not entirely satisfied with the full-width glyphs: they have glue between. If one looks closely at his utf8test-1.pdf test file, you can see that three glyphs on one line are not as wide as three glyphs on another line. I find it a little disturbing to read. OTOH his other document, bcac.pdf, shows at the end that it is somehow possible to achieve this. But bcac.pdf shows another problem in PDF: a lot of glyphs (in Latin script) look in evince like they have 1pt or 2pt on the right just "chopped off". And something else that I have noted throughout viewing PDF results of Omega/Lambda files: evince showed almost always some strange output. The title on the first page of Robert Wright's file start from the right margin (?!). Japanese texts that don't have fonts embbedded, will show up as gibberish since I don't have the exact same Adobe fonts as the author. It also doesn't help that one of the original authors has broken with Omega and announced that he would start with another TeX flavour. So switching to Omega/Lambda seems for me, currently, a bit too soon, especially when it concerns CJK, because the font process looks exactly the same as on LaTeX, with an extra stage of OVP fonts. XeTeX looks even more promising. It even supports OpenType features as you told, so things like GSUB for vertical CJK writing might be somewhat interesting. It seems like I will get grey hair at a young age. ;D > > Two things are needed though: an updated Unicode.sfd file in > > freetype1-tools (I just filed a bug report), and a DFSG-free font that > > covers most scripts. > > Why do you want a single font? Given that one anyway needs some sort of > mark up to switch hyphenation patterns for different languages, it is > quite easy to integrate font switches there as well. True. It might be best to use separate fonts. "Non multa sed multum", as I was always told. ;) Cheers Danai SAE-HAN 韓達耐 -- 題目:《宮詞》 作者:宋徽宗〔趙佶〕（１０８２－１１３５） 新樣梳妝巧畫眉，窄衣纖體最相宜； 一時趨向多情遠，小閣幽窗靜弈棋。
