What is needed is something like: http://www.ctan.org/tex-archive/macros/xetex/latex/fontwrap/
<http://www.ctan.org/tex-archive/macros/xetex/latex/fontwrap/>So that fonts are *autoselected* based on UNICODE range unless otherwise overridden. fontwrap package for XeLaTex ---------------------------- Michiel Kamermans, june 2008 Available under GPL What is it? -------------- fontwrap is a package for XeLaTeX to automatically wrap multilingual utf8 encoded text in approriate font tags, issuing fontspec commands whenever text switches from one unicode block to another. As most of us use TeX in order to separate visual styling from the actual document as much as possible, this package removes the need to add fonttags all over the place when writing multilingual documents. On Tue, Nov 9, 2010 at 9:02 AM, Walter <walter.stan...@gmail.com> wrote: > Daron, > > Thanks so much for your detailed reply! > > I was very inspired by your screenshot and output PDF, which have me > hope that a decent solution *is*, in fact, possible! > > I had no idea that it was possible to set up character styles with > font associations. This seems like the right sort of solution. > > First an update. Prior to your reply I did already try upgrading to > LyX 2.0 alpha6 as recommended, however it did not change my situation > very much as I had by that stage already set up TeX codeboxes before > and after every piece of non-English text to specifically change the > font. That solution 'works', except that it makes the document > annoying to edit, and it's a pain when I want to, say, change the font > used for Chinese right throughout the document. (A job for sed!) Also > in upgrading, I had failed to see the the 'Document > Settings > > Output > Use XeTeX' checkbox, which is great and seems to have a > follow-on effect on modifying various elements elsewhere within the > Document Settings dialog ... so thanks for pointing that one out. > > > If you can get the module to be recognized in LyX > > In attempting to replicate your solution, I have had problems > specifically with this part until I realised a full > start/reconfigure/stop/start cycle was necessary to make the new > module visible. I think I had assumed incorrectly that stop/start > would have the same effect. > > The result: finally, I did succeed in getting your solution to work. > Hooray! Thank you so much. If you are ever in Los > Angeles/London/Kunming/Sydney (my present four homes) and wish to have > a free 'adult beverage' (in deference to the LyX customization > document), then you are most welcome. > > Now, I will try to give something back. > > *ahem* Thoughts for the future, or "Dear LyX Community": > > It seems to me that the primary stated goal of LyX is to combine the > flexibility of TeX with the ease of use of a GUI. > > Right now, 'ease of use' would not describe accurately the process > required to achieve a relatively pleasant process for editing > multilingual documents. > > Whilst the recent addition of built-in XeTeX support to LyX 2.0 is a > huge step in the right direction, at present the website's claim of > "No more endless tinkering with formatting details, 'finger painting' > font attributes" is somewhat inflated when it comes to this sort of > situation. > > In effect, even with Daron's comparatively elegant solution, one must > still essentially finger paint structures corresponding to fonts, and > only after significant setup to enable the process. > > IMHO, the best possible solution for documents with heavy use of > multilingual text would seem to be a combination of this type of > solution (character styles with font associations), and automatic > unicode block-based input classification. This would enable LyX to > automatically change the default, available and/or recommended styles > based on what kind of character was input. > > Whilst there are a few obvious complications (Han characters - CJK - > which language?), if a user interface is provided by LyX that allows > the user to configure the matching process by defining unicode block > to character style and font associations, then 'just type'-level ease > of use could finally be achieved. Furthermore, by detecting the > available unicode glyphs in each font, intelligent font suggestions > could be made when defining a new, unicode block-associated character > style. > > Please find a suggested UI mockup attached. > > (Probably, the "language" concept within Daron's module should be > dropped in favour of "character style" with unicode block titles as > default nomenclature) > > What do others think of this proposal? > > - Walter > -- --- - Learn to speak Cherokee: http://www.cherokeelessons.com/ - Cherokee Language Help BBS/Chat: http://www.cherokeelessons.com/phpBB3/viewforum.php?f=6 - Cherokee Lessons PDF made with: http://www.lyx.org/
