Lars Gullik Bjønnes wrote: > I think one year is too much, and we should try to make it happen > faster than that. (At least get basic support for Unicode in rather > fast and then use some time to refine it) > > I am currently looking hard at the ICU library, and even if I don't > like the api super well, I think this will reduce the effort we will > have to put in tremendously. > > I'd grateful if others also could have a look at ICU. > http://icu.sourceforge.net/
Will do. > One other options is to make LyX truly be a Qt app (I am not sure that > I really like this), and use the Unicode support therein) No! Qt is very good for all sorts of things so long as you don't care about "beautiful code". Clearly, as a bunch, we care too much about "beautiful code", but I think that going all the way to Qt will kill us:) Have a look at ftp://ftp.u-aizu.ac.jp/pub/tex/cjk-lyx/qt/CJK-LyX-qt-1.3.6-1.patch to see what we need to do to get full CJK support in LyX (implies full Unicode support). Admittedly, the patch is a kludge and we'd do some stuff differently. However, the fact that the whole thing is only 400kB means we must be doing something right. > One thing that must happen together withthe Unicode work is to change > our on-disk encoding to be UTF-8. Right. But this just needs some UCS-4 <-> UTF-8 converters. It doesn't mean we have to ditch std::basic_string<>. > Of course bug fixes and small nifty new features will be allowed in. > But nothing major. 100% agree. Other major features should be developed in a branch and merged back in only when they're mature. -- Angus
